Final_Project.ipynb

{
  "cells": [
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "F9pPoIQgApH9",
        "outputId": "0d115b0b-eab7-48d2-c428-1665147969d9"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Mounted at /content/gdrive\n"
          ]
        }
      ],
      "source": [
        "from google.colab import drive\n",
        "drive.mount('/content/gdrive')"
      ],
      "id": "F9pPoIQgApH9"
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 110,
          "resources": {
            "http://localhost:8080/nbextensions/google.colab/files.js": {
              "data": "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",
              "headers": [
                [
                  "content-type",
                  "application/javascript"
                ]
              ],
              "ok": true,
              "status": 200,
              "status_text": ""
            }
          }
        },
        "id": "Em6ckbzICRGp",
        "outputId": "6c6266f7-27ce-4bba-c09a-ef9518f2fbe5"
      },
      "outputs": [
        {
          "data": {
            "text/html": [
              "\n",
              "     <input type=\"file\" id=\"files-86f23d2f-fd79-420a-9cfe-946f7d350adf\" name=\"files[]\" multiple disabled\n",
              "        style=\"border:none\" />\n",
              "     <output id=\"result-86f23d2f-fd79-420a-9cfe-946f7d350adf\">\n",
              "      Upload widget is only available when the cell has been executed in the\n",
              "      current browser session. Please rerun this cell to enable.\n",
              "      </output>\n",
              "      <script src=\"/nbextensions/google.colab/files.js\"></script> "
            ],
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Saving kaggle (1).json to kaggle (1).json\n"
          ]
        },
        {
          "data": {
            "text/plain": [
              "{'kaggle (1).json': b'{\"username\":\"gauravsingh37\",\"key\":\"aa555d8d832f7e239a366aae66d77d27\"}'}"
            ]
          },
          "execution_count": 2,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "from google.colab import files\n",
        "\n",
        "files.upload() #this will prompt you to upload the kaggle.json"
      ],
      "id": "Em6ckbzICRGp"
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "X2Yl6a7vCWxG",
        "outputId": "e92d4900-358b-4044-e28c-2b7f107dbaa8"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "-rw-r--r-- 1 root root 69 Aug 19 01:46 kaggle.json\n",
            "/content\n",
            "ref                                                             title                                           size  lastUpdated          downloadCount  voteCount  usabilityRating  \n",
            "--------------------------------------------------------------  ---------------------------------------------  -----  -------------------  -------------  ---------  ---------------  \n",
            "ruchi798/data-science-job-salaries                              Data Science Job Salaries                        7KB  2022-06-15 08:59:12          21164        651  1.0              \n",
            "josephcheng123456/olympic-historical-dataset-from-olympediaorg  Olympic Historical Dataset From Olympedia.org   25MB  2022-08-18 14:18:14            665         27  1.0              \n",
            "deepcontractor/mcdonalds-india-menu-nutrition-facts             McDonald's India : Menu Nutrition Dataset        5KB  2022-07-28 17:08:52           1697         90  0.9411765        \n",
            "gpreda/life-expectancy-in-europe-by-age-and-sex                 Life Expectancy in Europe by Age and Sex         4MB  2022-08-12 15:28:01            564         28  1.0              \n",
            "arianazmoudeh/airbnbopendata                                    Airbnb Open Data                                10MB  2022-08-01 15:58:10           1230         40  0.9705882        \n",
            "nadzmiagthomas/australia-data-science-jobs                      Australia Data Science Jobs                      2MB  2022-08-09 11:48:07            337         24  0.9705882        \n",
            "infamouscoder/dataset-netflix-shows                             Dataset: NetFlix Shows                           1MB  2022-08-11 05:05:41            541         37  1.0              \n",
            "himanshunakrani/student-study-hours                             Student Study Hours                             276B  2022-07-20 13:17:29           2209         75  1.0              \n",
            "nancyalaswad90/diamonds-prices                                  Diamonds Prices                                711KB  2022-07-09 14:59:21           2527        108  1.0              \n",
            "zzettrkalpakbal/full-filled-brain-stroke-dataset                Brain stroke prediction dataset                 52KB  2022-07-16 09:57:08           2577         72  0.9705882        \n",
            "reihanenamdari/breast-cancer                                    Breast Cancer                                   43KB  2022-08-08 19:25:55            463         31  1.0              \n",
            "dansbecker/melbourne-housing-snapshot                           Melbourne Housing Snapshot                     451KB  2018-06-05 12:52:24          93305       1125  0.7058824        \n",
            "jpmiller/employee-attrition-for-healthcare                      Employee Attrition for Healthcare               59KB  2022-08-10 22:34:29            562         37  0.9705882        \n",
            "ramjasmaurya/commonwealth-games-2022                            Commonwealth Games 2022                         80KB  2022-08-09 03:00:59            371         27  1.0              \n",
            "jillanisofttech/brain-stroke-dataset                            Brain Stroke Dataset                            47KB  2022-08-04 18:02:56            975         34  0.9705882        \n",
            "faryarmemon/usa-housing-market-factors                          U.S. Housing Market Factors                     32KB  2022-08-03 02:19:31            580         38  1.0              \n",
            "surajjha101/countries-olympics-medals-since-1896                Countries Olympics Medals since 1896             4KB  2022-07-27 06:39:06            983         54  1.0              \n",
            "datasnaek/youtube-new                                           Trending YouTube Video Statistics              201MB  2019-06-03 00:56:47         181658       4637  0.7941176        \n",
            "zynicide/wine-reviews                                           Wine Reviews                                    51MB  2017-11-27 17:08:04         164654       3353  0.7941176        \n",
            "erqizhou/students-data-analysis                                 Students Data Analysis                           2KB  2022-07-20 03:54:13           1165         38  1.0              \n"
          ]
        }
      ],
      "source": [
        "!ls -lha kaggle.json\n",
        "!pip install -q kaggle\n",
        "!mkdir -p ~/.kaggle\n",
        "!cp kaggle.json ~/.kaggle/\n",
        "!chmod 600 /root/.kaggle/kaggle.json\n",
        "!pwd\n",
        "!kaggle datasets list"
      ],
      "id": "X2Yl6a7vCWxG"
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "DUOk-JbTCW4P",
        "outputId": "22a442e9-b23c-4957-a1ac-9611b23da5d1"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Downloading facial-expression-dataset-image-folders-fer2013.zip to /content\n",
            " 88% 57.0M/65.1M [00:00<00:00, 96.9MB/s]\n",
            "100% 65.1M/65.1M [00:00<00:00, 94.6MB/s]\n"
          ]
        }
      ],
      "source": [
        "!kaggle datasets download -d astraszab/facial-expression-dataset-image-folders-fer2013"
      ],
      "id": "DUOk-JbTCW4P"
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Z_tiH6quKu4-",
        "outputId": "5bf6b8c2-971b-4c05-b8bb-0df8238db584"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "\u001b[1;30;43mStreaming output truncated to the last 5000 lines.\u001b[0m\n",
            "  inflating: data/train/6/28451.png  \n",
            "  inflating: data/train/6/28452.png  \n",
            "  inflating: data/train/6/28467.png  \n",
            "  inflating: data/train/6/28471.png  \n",
            "  inflating: data/train/6/28472.png  \n",
            "  inflating: data/train/6/28478.png  \n",
            "  inflating: data/train/6/2848.png   \n",
            "  inflating: data/train/6/28480.png  \n",
            "  inflating: data/train/6/28484.png  \n",
            "  inflating: data/train/6/28485.png  \n",
            "  inflating: data/train/6/28493.png  \n",
            "  inflating: data/train/6/28499.png  \n",
            "  inflating: data/train/6/28507.png  \n",
            "  inflating: data/train/6/28517.png  \n",
            "  inflating: data/train/6/28519.png  \n",
            "  inflating: data/train/6/28520.png  \n",
            "  inflating: data/train/6/28524.png  \n",
            "  inflating: data/train/6/28530.png  \n",
            "  inflating: data/train/6/28531.png  \n",
            "  inflating: data/train/6/28535.png  \n",
            "  inflating: data/train/6/28537.png  \n",
            "  inflating: data/train/6/28539.png  \n",
            "  inflating: data/train/6/28540.png  \n",
            "  inflating: data/train/6/28545.png  \n",
            "  inflating: data/train/6/28546.png  \n",
            "  inflating: data/train/6/28549.png  \n",
            "  inflating: data/train/6/28550.png  \n",
            "  inflating: data/train/6/28551.png  \n",
            "  inflating: data/train/6/2856.png   \n",
            "  inflating: data/train/6/28569.png  \n",
            "  inflating: data/train/6/28570.png  \n",
            "  inflating: data/train/6/28580.png  \n",
            "  inflating: data/train/6/28586.png  \n",
            "  inflating: data/train/6/28587.png  \n",
            "  inflating: data/train/6/28588.png  \n",
            "  inflating: data/train/6/28599.png  \n",
            "  inflating: data/train/6/286.png    \n",
            "  inflating: data/train/6/28602.png  \n",
            "  inflating: data/train/6/28609.png  \n",
            "  inflating: data/train/6/28617.png  \n",
            "  inflating: data/train/6/28619.png  \n",
            "  inflating: data/train/6/28625.png  \n",
            "  inflating: data/train/6/28626.png  \n",
            "  inflating: data/train/6/28630.png  \n",
            "  inflating: data/train/6/28637.png  \n",
            "  inflating: data/train/6/28642.png  \n",
            "  inflating: data/train/6/2865.png   \n",
            "  inflating: data/train/6/28652.png  \n",
            "  inflating: data/train/6/28664.png  \n",
            "  inflating: data/train/6/28672.png  \n",
            "  inflating: data/train/6/28679.png  \n",
            "  inflating: data/train/6/28681.png  \n",
            "  inflating: data/train/6/28693.png  \n",
            "  inflating: data/train/6/28699.png  \n",
            "  inflating: data/train/6/2870.png   \n",
            "  inflating: data/train/6/2871.png   \n",
            "  inflating: data/train/6/2873.png   \n",
            "  inflating: data/train/6/2876.png   \n",
            "  inflating: data/train/6/2882.png   \n",
            "  inflating: data/train/6/2887.png   \n",
            "  inflating: data/train/6/2888.png   \n",
            "  inflating: data/train/6/2893.png   \n",
            "  inflating: data/train/6/2895.png   \n",
            "  inflating: data/train/6/290.png    \n",
            "  inflating: data/train/6/2901.png   \n",
            "  inflating: data/train/6/2914.png   \n",
            "  inflating: data/train/6/2923.png   \n",
            "  inflating: data/train/6/2926.png   \n",
            "  inflating: data/train/6/2932.png   \n",
            "  inflating: data/train/6/2935.png   \n",
            "  inflating: data/train/6/2937.png   \n",
            "  inflating: data/train/6/2953.png   \n",
            "  inflating: data/train/6/2955.png   \n",
            "  inflating: data/train/6/2957.png   \n",
            "  inflating: data/train/6/2962.png   \n",
            "  inflating: data/train/6/2967.png   \n",
            "  inflating: data/train/6/2973.png   \n",
            "  inflating: data/train/6/2974.png   \n",
            "  inflating: data/train/6/2977.png   \n",
            "  inflating: data/train/6/2978.png   \n",
            "  inflating: data/train/6/2984.png   \n",
            "  inflating: data/train/6/2986.png   \n",
            "  inflating: data/train/6/2995.png   \n",
            "  inflating: data/train/6/2998.png   \n",
            "  inflating: data/train/6/3001.png   \n",
            "  inflating: data/train/6/3006.png   \n",
            "  inflating: data/train/6/3014.png   \n",
            "  inflating: data/train/6/3018.png   \n",
            "  inflating: data/train/6/3025.png   \n",
            "  inflating: data/train/6/3041.png   \n",
            "  inflating: data/train/6/3049.png   \n",
            "  inflating: data/train/6/3050.png   \n",
            "  inflating: data/train/6/3051.png   \n",
            "  inflating: data/train/6/3055.png   \n",
            "  inflating: data/train/6/3058.png   \n",
            "  inflating: data/train/6/3061.png   \n",
            "  inflating: data/train/6/3064.png   \n",
            "  inflating: data/train/6/3066.png   \n",
            "  inflating: data/train/6/3071.png   \n",
            "  inflating: data/train/6/3072.png   \n",
            "  inflating: data/train/6/3076.png   \n",
            "  inflating: data/train/6/3079.png   \n",
            "  inflating: data/train/6/3081.png   \n",
            "  inflating: data/train/6/3093.png   \n",
            "  inflating: data/train/6/3095.png   \n",
            "  inflating: data/train/6/3098.png   \n",
            "  inflating: data/train/6/31.png     \n",
            "  inflating: data/train/6/3109.png   \n",
            "  inflating: data/train/6/311.png    \n",
            "  inflating: data/train/6/3114.png   \n",
            "  inflating: data/train/6/3115.png   \n",
            "  inflating: data/train/6/3123.png   \n",
            "  inflating: data/train/6/3127.png   \n",
            "  inflating: data/train/6/3128.png   \n",
            "  inflating: data/train/6/3140.png   \n",
            "  inflating: data/train/6/3148.png   \n",
            "  inflating: data/train/6/3153.png   \n",
            "  inflating: data/train/6/3155.png   \n",
            "  inflating: data/train/6/3156.png   \n",
            "  inflating: data/train/6/3161.png   \n",
            "  inflating: data/train/6/3165.png   \n",
            "  inflating: data/train/6/3166.png   \n",
            "  inflating: data/train/6/3172.png   \n",
            "  inflating: data/train/6/3173.png   \n",
            "  inflating: data/train/6/3174.png   \n",
            "  inflating: data/train/6/3177.png   \n",
            "  inflating: data/train/6/3179.png   \n",
            "  inflating: data/train/6/3190.png   \n",
            "  inflating: data/train/6/3191.png   \n",
            "  inflating: data/train/6/3198.png   \n",
            "  inflating: data/train/6/320.png    \n",
            "  inflating: data/train/6/3204.png   \n",
            "  inflating: data/train/6/3221.png   \n",
            "  inflating: data/train/6/3228.png   \n",
            "  inflating: data/train/6/3231.png   \n",
            "  inflating: data/train/6/3236.png   \n",
            "  inflating: data/train/6/3244.png   \n",
            "  inflating: data/train/6/3247.png   \n",
            "  inflating: data/train/6/3258.png   \n",
            "  inflating: data/train/6/3266.png   \n",
            "  inflating: data/train/6/3268.png   \n",
            "  inflating: data/train/6/3272.png   \n",
            "  inflating: data/train/6/3276.png   \n",
            "  inflating: data/train/6/3280.png   \n",
            "  inflating: data/train/6/3288.png   \n",
            "  inflating: data/train/6/3290.png   \n",
            "  inflating: data/train/6/3305.png   \n",
            "  inflating: data/train/6/3307.png   \n",
            "  inflating: data/train/6/3312.png   \n",
            "  inflating: data/train/6/3321.png   \n",
            "  inflating: data/train/6/3323.png   \n",
            "  inflating: data/train/6/3325.png   \n",
            "  inflating: data/train/6/3329.png   \n",
            "  inflating: data/train/6/3333.png   \n",
            "  inflating: data/train/6/3336.png   \n",
            "  inflating: data/train/6/3340.png   \n",
            "  inflating: data/train/6/3346.png   \n",
            "  inflating: data/train/6/3353.png   \n",
            "  inflating: data/train/6/3355.png   \n",
            "  inflating: data/train/6/3362.png   \n",
            "  inflating: data/train/6/337.png    \n",
            "  inflating: data/train/6/3371.png   \n",
            "  inflating: data/train/6/3375.png   \n",
            "  inflating: data/train/6/3379.png   \n",
            "  inflating: data/train/6/3380.png   \n",
            "  inflating: data/train/6/3383.png   \n",
            "  inflating: data/train/6/3384.png   \n",
            "  inflating: data/train/6/3395.png   \n",
            "  inflating: data/train/6/3396.png   \n",
            "  inflating: data/train/6/3401.png   \n",
            "  inflating: data/train/6/3404.png   \n",
            "  inflating: data/train/6/3406.png   \n",
            "  inflating: data/train/6/3419.png   \n",
            "  inflating: data/train/6/3424.png   \n",
            "  inflating: data/train/6/3441.png   \n",
            "  inflating: data/train/6/3444.png   \n",
            "  inflating: data/train/6/346.png    \n",
            "  inflating: data/train/6/347.png    \n",
            "  inflating: data/train/6/3476.png   \n",
            "  inflating: data/train/6/3485.png   \n",
            "  inflating: data/train/6/3494.png   \n",
            "  inflating: data/train/6/3497.png   \n",
            "  inflating: data/train/6/3498.png   \n",
            "  inflating: data/train/6/35.png     \n",
            "  inflating: data/train/6/3500.png   \n",
            "  inflating: data/train/6/3502.png   \n",
            "  inflating: data/train/6/3505.png   \n",
            "  inflating: data/train/6/3510.png   \n",
            "  inflating: data/train/6/3511.png   \n",
            "  inflating: data/train/6/3527.png   \n",
            "  inflating: data/train/6/3535.png   \n",
            "  inflating: data/train/6/3537.png   \n",
            "  inflating: data/train/6/3545.png   \n",
            "  inflating: data/train/6/3557.png   \n",
            "  inflating: data/train/6/3560.png   \n",
            "  inflating: data/train/6/3566.png   \n",
            "  inflating: data/train/6/3568.png   \n",
            "  inflating: data/train/6/3571.png   \n",
            "  inflating: data/train/6/3575.png   \n",
            "  inflating: data/train/6/3581.png   \n",
            "  inflating: data/train/6/3588.png   \n",
            "  inflating: data/train/6/3589.png   \n",
            "  inflating: data/train/6/3592.png   \n",
            "  inflating: data/train/6/3597.png   \n",
            "  inflating: data/train/6/36.png     \n",
            "  inflating: data/train/6/3600.png   \n",
            "  inflating: data/train/6/3610.png   \n",
            "  inflating: data/train/6/3614.png   \n",
            "  inflating: data/train/6/3617.png   \n",
            "  inflating: data/train/6/3619.png   \n",
            "  inflating: data/train/6/3623.png   \n",
            "  inflating: data/train/6/363.png    \n",
            "  inflating: data/train/6/3633.png   \n",
            "  inflating: data/train/6/3656.png   \n",
            "  inflating: data/train/6/3665.png   \n",
            "  inflating: data/train/6/3672.png   \n",
            "  inflating: data/train/6/3674.png   \n",
            "  inflating: data/train/6/3676.png   \n",
            "  inflating: data/train/6/3677.png   \n",
            "  inflating: data/train/6/3681.png   \n",
            "  inflating: data/train/6/3688.png   \n",
            "  inflating: data/train/6/3696.png   \n",
            "  inflating: data/train/6/3697.png   \n",
            "  inflating: data/train/6/37.png     \n",
            "  inflating: data/train/6/3700.png   \n",
            "  inflating: data/train/6/3706.png   \n",
            "  inflating: data/train/6/3711.png   \n",
            "  inflating: data/train/6/3715.png   \n",
            "  inflating: data/train/6/3718.png   \n",
            "  inflating: data/train/6/3726.png   \n",
            "  inflating: data/train/6/3729.png   \n",
            "  inflating: data/train/6/373.png    \n",
            "  inflating: data/train/6/3732.png   \n",
            "  inflating: data/train/6/3733.png   \n",
            "  inflating: data/train/6/3735.png   \n",
            "  inflating: data/train/6/3745.png   \n",
            "  inflating: data/train/6/376.png    \n",
            "  inflating: data/train/6/3761.png   \n",
            "  inflating: data/train/6/3771.png   \n",
            "  inflating: data/train/6/3774.png   \n",
            "  inflating: data/train/6/3779.png   \n",
            "  inflating: data/train/6/3790.png   \n",
            "  inflating: data/train/6/3792.png   \n",
            "  inflating: data/train/6/3795.png   \n",
            "  inflating: data/train/6/3798.png   \n",
            "  inflating: data/train/6/3803.png   \n",
            "  inflating: data/train/6/3805.png   \n",
            "  inflating: data/train/6/3806.png   \n",
            "  inflating: data/train/6/3809.png   \n",
            "  inflating: data/train/6/3816.png   \n",
            "  inflating: data/train/6/382.png    \n",
            "  inflating: data/train/6/3830.png   \n",
            "  inflating: data/train/6/3831.png   \n",
            "  inflating: data/train/6/3832.png   \n",
            "  inflating: data/train/6/3844.png   \n",
            "  inflating: data/train/6/3849.png   \n",
            "  inflating: data/train/6/3856.png   \n",
            "  inflating: data/train/6/3871.png   \n",
            "  inflating: data/train/6/3886.png   \n",
            "  inflating: data/train/6/3890.png   \n",
            "  inflating: data/train/6/3898.png   \n",
            "  inflating: data/train/6/3899.png   \n",
            "  inflating: data/train/6/3907.png   \n",
            "  inflating: data/train/6/3909.png   \n",
            "  inflating: data/train/6/391.png    \n",
            "  inflating: data/train/6/3912.png   \n",
            "  inflating: data/train/6/3924.png   \n",
            "  inflating: data/train/6/3929.png   \n",
            "  inflating: data/train/6/3932.png   \n",
            "  inflating: data/train/6/3934.png   \n",
            "  inflating: data/train/6/3938.png   \n",
            "  inflating: data/train/6/3943.png   \n",
            "  inflating: data/train/6/3944.png   \n",
            "  inflating: data/train/6/3946.png   \n",
            "  inflating: data/train/6/3948.png   \n",
            "  inflating: data/train/6/395.png    \n",
            "  inflating: data/train/6/3953.png   \n",
            "  inflating: data/train/6/3954.png   \n",
            "  inflating: data/train/6/3955.png   \n",
            "  inflating: data/train/6/396.png    \n",
            "  inflating: data/train/6/397.png    \n",
            "  inflating: data/train/6/3980.png   \n",
            "  inflating: data/train/6/3986.png   \n",
            "  inflating: data/train/6/399.png    \n",
            "  inflating: data/train/6/3996.png   \n",
            "  inflating: data/train/6/3997.png   \n",
            "  inflating: data/train/6/4.png      \n",
            "  inflating: data/train/6/4003.png   \n",
            "  inflating: data/train/6/4007.png   \n",
            "  inflating: data/train/6/4026.png   \n",
            "  inflating: data/train/6/405.png    \n",
            "  inflating: data/train/6/4057.png   \n",
            "  inflating: data/train/6/4061.png   \n",
            "  inflating: data/train/6/4062.png   \n",
            "  inflating: data/train/6/4069.png   \n",
            "  inflating: data/train/6/4081.png   \n",
            "  inflating: data/train/6/4083.png   \n",
            "  inflating: data/train/6/4087.png   \n",
            "  inflating: data/train/6/4092.png   \n",
            "  inflating: data/train/6/410.png    \n",
            "  inflating: data/train/6/4105.png   \n",
            "  inflating: data/train/6/4122.png   \n",
            "  inflating: data/train/6/4124.png   \n",
            "  inflating: data/train/6/4127.png   \n",
            "  inflating: data/train/6/4132.png   \n",
            "  inflating: data/train/6/4137.png   \n",
            "  inflating: data/train/6/4138.png   \n",
            "  inflating: data/train/6/4144.png   \n",
            "  inflating: data/train/6/4145.png   \n",
            "  inflating: data/train/6/415.png    \n",
            "  inflating: data/train/6/4151.png   \n",
            "  inflating: data/train/6/4160.png   \n",
            "  inflating: data/train/6/4161.png   \n",
            "  inflating: data/train/6/4162.png   \n",
            "  inflating: data/train/6/4166.png   \n",
            "  inflating: data/train/6/4178.png   \n",
            "  inflating: data/train/6/4184.png   \n",
            "  inflating: data/train/6/4187.png   \n",
            "  inflating: data/train/6/419.png    \n",
            "  inflating: data/train/6/4190.png   \n",
            "  inflating: data/train/6/4194.png   \n",
            "  inflating: data/train/6/4195.png   \n",
            "  inflating: data/train/6/4202.png   \n",
            "  inflating: data/train/6/4211.png   \n",
            "  inflating: data/train/6/4213.png   \n",
            "  inflating: data/train/6/4214.png   \n",
            "  inflating: data/train/6/4218.png   \n",
            "  inflating: data/train/6/4219.png   \n",
            "  inflating: data/train/6/4223.png   \n",
            "  inflating: data/train/6/4228.png   \n",
            "  inflating: data/train/6/423.png    \n",
            "  inflating: data/train/6/4231.png   \n",
            "  inflating: data/train/6/4241.png   \n",
            "  inflating: data/train/6/4242.png   \n",
            "  inflating: data/train/6/4247.png   \n",
            "  inflating: data/train/6/4253.png   \n",
            "  inflating: data/train/6/4258.png   \n",
            "  inflating: data/train/6/4263.png   \n",
            "  inflating: data/train/6/4266.png   \n",
            "  inflating: data/train/6/4268.png   \n",
            "  inflating: data/train/6/4270.png   \n",
            "  inflating: data/train/6/4279.png   \n",
            "  inflating: data/train/6/4282.png   \n",
            "  inflating: data/train/6/4286.png   \n",
            "  inflating: data/train/6/4302.png   \n",
            "  inflating: data/train/6/4307.png   \n",
            "  inflating: data/train/6/4309.png   \n",
            "  inflating: data/train/6/4314.png   \n",
            "  inflating: data/train/6/4320.png   \n",
            "  inflating: data/train/6/4325.png   \n",
            "  inflating: data/train/6/4328.png   \n",
            "  inflating: data/train/6/4339.png   \n",
            "  inflating: data/train/6/4346.png   \n",
            "  inflating: data/train/6/4348.png   \n",
            "  inflating: data/train/6/4350.png   \n",
            "  inflating: data/train/6/4354.png   \n",
            "  inflating: data/train/6/4361.png   \n",
            "  inflating: data/train/6/4372.png   \n",
            "  inflating: data/train/6/4380.png   \n",
            "  inflating: data/train/6/4383.png   \n",
            "  inflating: data/train/6/4386.png   \n",
            "  inflating: data/train/6/4387.png   \n",
            "  inflating: data/train/6/4388.png   \n",
            "  inflating: data/train/6/4394.png   \n",
            "  inflating: data/train/6/44.png     \n",
            "  inflating: data/train/6/4409.png   \n",
            "  inflating: data/train/6/4413.png   \n",
            "  inflating: data/train/6/442.png    \n",
            "  inflating: data/train/6/4428.png   \n",
            "  inflating: data/train/6/4437.png   \n",
            "  inflating: data/train/6/4443.png   \n",
            "  inflating: data/train/6/4449.png   \n",
            "  inflating: data/train/6/4463.png   \n",
            "  inflating: data/train/6/4465.png   \n",
            "  inflating: data/train/6/449.png    \n",
            "  inflating: data/train/6/4498.png   \n",
            "  inflating: data/train/6/4499.png   \n",
            "  inflating: data/train/6/4519.png   \n",
            "  inflating: data/train/6/4521.png   \n",
            "  inflating: data/train/6/4524.png   \n",
            "  inflating: data/train/6/4533.png   \n",
            "  inflating: data/train/6/4535.png   \n",
            "  inflating: data/train/6/4550.png   \n",
            "  inflating: data/train/6/4551.png   \n",
            "  inflating: data/train/6/4552.png   \n",
            "  inflating: data/train/6/4553.png   \n",
            "  inflating: data/train/6/4561.png   \n",
            "  inflating: data/train/6/4563.png   \n",
            "  inflating: data/train/6/4573.png   \n",
            "  inflating: data/train/6/4574.png   \n",
            "  inflating: data/train/6/4584.png   \n",
            "  inflating: data/train/6/459.png    \n",
            "  inflating: data/train/6/4591.png   \n",
            "  inflating: data/train/6/4599.png   \n",
            "  inflating: data/train/6/461.png    \n",
            "  inflating: data/train/6/4625.png   \n",
            "  inflating: data/train/6/467.png    \n",
            "  inflating: data/train/6/4671.png   \n",
            "  inflating: data/train/6/4672.png   \n",
            "  inflating: data/train/6/468.png    \n",
            "  inflating: data/train/6/4691.png   \n",
            "  inflating: data/train/6/4699.png   \n",
            "  inflating: data/train/6/47.png     \n",
            "  inflating: data/train/6/4702.png   \n",
            "  inflating: data/train/6/4704.png   \n",
            "  inflating: data/train/6/4709.png   \n",
            "  inflating: data/train/6/4711.png   \n",
            "  inflating: data/train/6/4718.png   \n",
            "  inflating: data/train/6/4722.png   \n",
            "  inflating: data/train/6/4730.png   \n",
            "  inflating: data/train/6/4735.png   \n",
            "  inflating: data/train/6/4744.png   \n",
            "  inflating: data/train/6/4745.png   \n",
            "  inflating: data/train/6/4748.png   \n",
            "  inflating: data/train/6/4752.png   \n",
            "  inflating: data/train/6/4760.png   \n",
            "  inflating: data/train/6/4768.png   \n",
            "  inflating: data/train/6/4769.png   \n",
            "  inflating: data/train/6/4770.png   \n",
            "  inflating: data/train/6/4777.png   \n",
            "  inflating: data/train/6/4781.png   \n",
            "  inflating: data/train/6/4787.png   \n",
            "  inflating: data/train/6/4789.png   \n",
            "  inflating: data/train/6/4790.png   \n",
            "  inflating: data/train/6/4798.png   \n",
            "  inflating: data/train/6/4799.png   \n",
            "  inflating: data/train/6/4802.png   \n",
            "  inflating: data/train/6/482.png    \n",
            "  inflating: data/train/6/4820.png   \n",
            "  inflating: data/train/6/4822.png   \n",
            "  inflating: data/train/6/4825.png   \n",
            "  inflating: data/train/6/4829.png   \n",
            "  inflating: data/train/6/4836.png   \n",
            "  inflating: data/train/6/4845.png   \n",
            "  inflating: data/train/6/4856.png   \n",
            "  inflating: data/train/6/4867.png   \n",
            "  inflating: data/train/6/4868.png   \n",
            "  inflating: data/train/6/4870.png   \n",
            "  inflating: data/train/6/4873.png   \n",
            "  inflating: data/train/6/4876.png   \n",
            "  inflating: data/train/6/4883.png   \n",
            "  inflating: data/train/6/4884.png   \n",
            "  inflating: data/train/6/4888.png   \n",
            "  inflating: data/train/6/4890.png   \n",
            "  inflating: data/train/6/4894.png   \n",
            "  inflating: data/train/6/4896.png   \n",
            "  inflating: data/train/6/4898.png   \n",
            "  inflating: data/train/6/490.png    \n",
            "  inflating: data/train/6/4903.png   \n",
            "  inflating: data/train/6/4906.png   \n",
            "  inflating: data/train/6/491.png    \n",
            "  inflating: data/train/6/4919.png   \n",
            "  inflating: data/train/6/4922.png   \n",
            "  inflating: data/train/6/4923.png   \n",
            "  inflating: data/train/6/4924.png   \n",
            "  inflating: data/train/6/493.png    \n",
            "  inflating: data/train/6/4933.png   \n",
            "  inflating: data/train/6/4935.png   \n",
            "  inflating: data/train/6/4953.png   \n",
            "  inflating: data/train/6/4960.png   \n",
            "  inflating: data/train/6/4968.png   \n",
            "  inflating: data/train/6/4974.png   \n",
            "  inflating: data/train/6/498.png    \n",
            "  inflating: data/train/6/4982.png   \n",
            "  inflating: data/train/6/4989.png   \n",
            "  inflating: data/train/6/4997.png   \n",
            "  inflating: data/train/6/4998.png   \n",
            "  inflating: data/train/6/5001.png   \n",
            "  inflating: data/train/6/5002.png   \n",
            "  inflating: data/train/6/501.png    \n",
            "  inflating: data/train/6/5013.png   \n",
            "  inflating: data/train/6/5015.png   \n",
            "  inflating: data/train/6/5017.png   \n",
            "  inflating: data/train/6/5018.png   \n",
            "  inflating: data/train/6/502.png    \n",
            "  inflating: data/train/6/5021.png   \n",
            "  inflating: data/train/6/503.png    \n",
            "  inflating: data/train/6/5034.png   \n",
            "  inflating: data/train/6/5044.png   \n",
            "  inflating: data/train/6/5046.png   \n",
            "  inflating: data/train/6/5049.png   \n",
            "  inflating: data/train/6/5051.png   \n",
            "  inflating: data/train/6/5053.png   \n",
            "  inflating: data/train/6/506.png    \n",
            "  inflating: data/train/6/5073.png   \n",
            "  inflating: data/train/6/5075.png   \n",
            "  inflating: data/train/6/508.png    \n",
            "  inflating: data/train/6/5087.png   \n",
            "  inflating: data/train/6/5099.png   \n",
            "  inflating: data/train/6/5105.png   \n",
            "  inflating: data/train/6/5110.png   \n",
            "  inflating: data/train/6/5116.png   \n",
            "  inflating: data/train/6/5119.png   \n",
            "  inflating: data/train/6/5136.png   \n",
            "  inflating: data/train/6/5139.png   \n",
            "  inflating: data/train/6/5143.png   \n",
            "  inflating: data/train/6/5157.png   \n",
            "  inflating: data/train/6/5161.png   \n",
            "  inflating: data/train/6/5170.png   \n",
            "  inflating: data/train/6/5174.png   \n",
            "  inflating: data/train/6/5186.png   \n",
            "  inflating: data/train/6/5193.png   \n",
            "  inflating: data/train/6/5194.png   \n",
            "  inflating: data/train/6/520.png    \n",
            "  inflating: data/train/6/5207.png   \n",
            "  inflating: data/train/6/5219.png   \n",
            "  inflating: data/train/6/522.png    \n",
            "  inflating: data/train/6/5225.png   \n",
            "  inflating: data/train/6/5226.png   \n",
            "  inflating: data/train/6/5227.png   \n",
            "  inflating: data/train/6/5230.png   \n",
            "  inflating: data/train/6/5231.png   \n",
            "  inflating: data/train/6/5248.png   \n",
            "  inflating: data/train/6/5255.png   \n",
            "  inflating: data/train/6/5258.png   \n",
            "  inflating: data/train/6/5260.png   \n",
            "  inflating: data/train/6/5267.png   \n",
            "  inflating: data/train/6/5272.png   \n",
            "  inflating: data/train/6/5275.png   \n",
            "  inflating: data/train/6/5292.png   \n",
            "  inflating: data/train/6/5297.png   \n",
            "  inflating: data/train/6/53.png     \n",
            "  inflating: data/train/6/5320.png   \n",
            "  inflating: data/train/6/5321.png   \n",
            "  inflating: data/train/6/5323.png   \n",
            "  inflating: data/train/6/5326.png   \n",
            "  inflating: data/train/6/5329.png   \n",
            "  inflating: data/train/6/5330.png   \n",
            "  inflating: data/train/6/5331.png   \n",
            "  inflating: data/train/6/5344.png   \n",
            "  inflating: data/train/6/5345.png   \n",
            "  inflating: data/train/6/535.png    \n",
            "  inflating: data/train/6/5353.png   \n",
            "  inflating: data/train/6/5358.png   \n",
            "  inflating: data/train/6/536.png    \n",
            "  inflating: data/train/6/5361.png   \n",
            "  inflating: data/train/6/5367.png   \n",
            "  inflating: data/train/6/5376.png   \n",
            "  inflating: data/train/6/5381.png   \n",
            "  inflating: data/train/6/5382.png   \n",
            "  inflating: data/train/6/5385.png   \n",
            "  inflating: data/train/6/5395.png   \n",
            "  inflating: data/train/6/5397.png   \n",
            "  inflating: data/train/6/5401.png   \n",
            "  inflating: data/train/6/5404.png   \n",
            "  inflating: data/train/6/5405.png   \n",
            "  inflating: data/train/6/5407.png   \n",
            "  inflating: data/train/6/541.png    \n",
            "  inflating: data/train/6/5411.png   \n",
            "  inflating: data/train/6/542.png    \n",
            "  inflating: data/train/6/5437.png   \n",
            "  inflating: data/train/6/544.png    \n",
            "  inflating: data/train/6/5444.png   \n",
            "  inflating: data/train/6/5446.png   \n",
            "  inflating: data/train/6/5447.png   \n",
            "  inflating: data/train/6/5449.png   \n",
            "  inflating: data/train/6/5454.png   \n",
            "  inflating: data/train/6/5457.png   \n",
            "  inflating: data/train/6/5465.png   \n",
            "  inflating: data/train/6/547.png    \n",
            "  inflating: data/train/6/5470.png   \n",
            "  inflating: data/train/6/5471.png   \n",
            "  inflating: data/train/6/5474.png   \n",
            "  inflating: data/train/6/5478.png   \n",
            "  inflating: data/train/6/5479.png   \n",
            "  inflating: data/train/6/5482.png   \n",
            "  inflating: data/train/6/5485.png   \n",
            "  inflating: data/train/6/5499.png   \n",
            "  inflating: data/train/6/5501.png   \n",
            "  inflating: data/train/6/5502.png   \n",
            "  inflating: data/train/6/551.png    \n",
            "  inflating: data/train/6/5510.png   \n",
            "  inflating: data/train/6/552.png    \n",
            "  inflating: data/train/6/5526.png   \n",
            "  inflating: data/train/6/5527.png   \n",
            "  inflating: data/train/6/5536.png   \n",
            "  inflating: data/train/6/5540.png   \n",
            "  inflating: data/train/6/5549.png   \n",
            "  inflating: data/train/6/555.png    \n",
            "  inflating: data/train/6/5552.png   \n",
            "  inflating: data/train/6/5556.png   \n",
            "  inflating: data/train/6/5561.png   \n",
            "  inflating: data/train/6/557.png    \n",
            "  inflating: data/train/6/5572.png   \n",
            "  inflating: data/train/6/5579.png   \n",
            "  inflating: data/train/6/5584.png   \n",
            "  inflating: data/train/6/5589.png   \n",
            "  inflating: data/train/6/5591.png   \n",
            "  inflating: data/train/6/5596.png   \n",
            "  inflating: data/train/6/5599.png   \n",
            "  inflating: data/train/6/5602.png   \n",
            "  inflating: data/train/6/5608.png   \n",
            "  inflating: data/train/6/5611.png   \n",
            "  inflating: data/train/6/5621.png   \n",
            "  inflating: data/train/6/5624.png   \n",
            "  inflating: data/train/6/5625.png   \n",
            "  inflating: data/train/6/5630.png   \n",
            "  inflating: data/train/6/5643.png   \n",
            "  inflating: data/train/6/5648.png   \n",
            "  inflating: data/train/6/565.png    \n",
            "  inflating: data/train/6/5651.png   \n",
            "  inflating: data/train/6/5652.png   \n",
            "  inflating: data/train/6/5660.png   \n",
            "  inflating: data/train/6/5669.png   \n",
            "  inflating: data/train/6/567.png    \n",
            "  inflating: data/train/6/5675.png   \n",
            "  inflating: data/train/6/5684.png   \n",
            "  inflating: data/train/6/5687.png   \n",
            "  inflating: data/train/6/5690.png   \n",
            "  inflating: data/train/6/5693.png   \n",
            "  inflating: data/train/6/5695.png   \n",
            "  inflating: data/train/6/5699.png   \n",
            "  inflating: data/train/6/570.png    \n",
            "  inflating: data/train/6/5708.png   \n",
            "  inflating: data/train/6/5713.png   \n",
            "  inflating: data/train/6/5715.png   \n",
            "  inflating: data/train/6/5716.png   \n",
            "  inflating: data/train/6/5724.png   \n",
            "  inflating: data/train/6/5734.png   \n",
            "  inflating: data/train/6/5745.png   \n",
            "  inflating: data/train/6/5748.png   \n",
            "  inflating: data/train/6/5749.png   \n",
            "  inflating: data/train/6/5754.png   \n",
            "  inflating: data/train/6/5775.png   \n",
            "  inflating: data/train/6/5784.png   \n",
            "  inflating: data/train/6/5792.png   \n",
            "  inflating: data/train/6/5811.png   \n",
            "  inflating: data/train/6/5813.png   \n",
            "  inflating: data/train/6/5817.png   \n",
            "  inflating: data/train/6/5818.png   \n",
            "  inflating: data/train/6/5831.png   \n",
            "  inflating: data/train/6/584.png    \n",
            "  inflating: data/train/6/5844.png   \n",
            "  inflating: data/train/6/5849.png   \n",
            "  inflating: data/train/6/5851.png   \n",
            "  inflating: data/train/6/5853.png   \n",
            "  inflating: data/train/6/5854.png   \n",
            "  inflating: data/train/6/5864.png   \n",
            "  inflating: data/train/6/5865.png   \n",
            "  inflating: data/train/6/5869.png   \n",
            "  inflating: data/train/6/5873.png   \n",
            "  inflating: data/train/6/5877.png   \n",
            "  inflating: data/train/6/5879.png   \n",
            "  inflating: data/train/6/5884.png   \n",
            "  inflating: data/train/6/5885.png   \n",
            "  inflating: data/train/6/5886.png   \n",
            "  inflating: data/train/6/5887.png   \n",
            "  inflating: data/train/6/5890.png   \n",
            "  inflating: data/train/6/5895.png   \n",
            "  inflating: data/train/6/5897.png   \n",
            "  inflating: data/train/6/5906.png   \n",
            "  inflating: data/train/6/5912.png   \n",
            "  inflating: data/train/6/5916.png   \n",
            "  inflating: data/train/6/592.png    \n",
            "  inflating: data/train/6/5920.png   \n",
            "  inflating: data/train/6/5927.png   \n",
            "  inflating: data/train/6/5931.png   \n",
            "  inflating: data/train/6/5945.png   \n",
            "  inflating: data/train/6/5953.png   \n",
            "  inflating: data/train/6/596.png    \n",
            "  inflating: data/train/6/5965.png   \n",
            "  inflating: data/train/6/5971.png   \n",
            "  inflating: data/train/6/5973.png   \n",
            "  inflating: data/train/6/5976.png   \n",
            "  inflating: data/train/6/5980.png   \n",
            "  inflating: data/train/6/5983.png   \n",
            "  inflating: data/train/6/5984.png   \n",
            "  inflating: data/train/6/5986.png   \n",
            "  inflating: data/train/6/5993.png   \n",
            "  inflating: data/train/6/5998.png   \n",
            "  inflating: data/train/6/60.png     \n",
            "  inflating: data/train/6/601.png    \n",
            "  inflating: data/train/6/6016.png   \n",
            "  inflating: data/train/6/6018.png   \n",
            "  inflating: data/train/6/602.png    \n",
            "  inflating: data/train/6/6027.png   \n",
            "  inflating: data/train/6/6034.png   \n",
            "  inflating: data/train/6/6038.png   \n",
            "  inflating: data/train/6/604.png    \n",
            "  inflating: data/train/6/6040.png   \n",
            "  inflating: data/train/6/605.png    \n",
            "  inflating: data/train/6/6056.png   \n",
            "  inflating: data/train/6/6065.png   \n",
            "  inflating: data/train/6/6069.png   \n",
            "  inflating: data/train/6/6078.png   \n",
            "  inflating: data/train/6/6093.png   \n",
            "  inflating: data/train/6/6102.png   \n",
            "  inflating: data/train/6/6103.png   \n",
            "  inflating: data/train/6/6143.png   \n",
            "  inflating: data/train/6/6148.png   \n",
            "  inflating: data/train/6/615.png    \n",
            "  inflating: data/train/6/6156.png   \n",
            "  inflating: data/train/6/6158.png   \n",
            "  inflating: data/train/6/6166.png   \n",
            "  inflating: data/train/6/6174.png   \n",
            "  inflating: data/train/6/6178.png   \n",
            "  inflating: data/train/6/6180.png   \n",
            "  inflating: data/train/6/619.png    \n",
            "  inflating: data/train/6/6192.png   \n",
            "  inflating: data/train/6/6193.png   \n",
            "  inflating: data/train/6/6195.png   \n",
            "  inflating: data/train/6/6196.png   \n",
            "  inflating: data/train/6/6203.png   \n",
            "  inflating: data/train/6/6205.png   \n",
            "  inflating: data/train/6/6213.png   \n",
            "  inflating: data/train/6/6221.png   \n",
            "  inflating: data/train/6/6234.png   \n",
            "  inflating: data/train/6/6236.png   \n",
            "  inflating: data/train/6/6238.png   \n",
            "  inflating: data/train/6/6243.png   \n",
            "  inflating: data/train/6/6255.png   \n",
            "  inflating: data/train/6/6266.png   \n",
            "  inflating: data/train/6/627.png    \n",
            "  inflating: data/train/6/6277.png   \n",
            "  inflating: data/train/6/6289.png   \n",
            "  inflating: data/train/6/6296.png   \n",
            "  inflating: data/train/6/6297.png   \n",
            "  inflating: data/train/6/630.png    \n",
            "  inflating: data/train/6/6315.png   \n",
            "  inflating: data/train/6/6325.png   \n",
            "  inflating: data/train/6/6326.png   \n",
            "  inflating: data/train/6/6335.png   \n",
            "  inflating: data/train/6/6340.png   \n",
            "  inflating: data/train/6/6343.png   \n",
            "  inflating: data/train/6/6348.png   \n",
            "  inflating: data/train/6/6351.png   \n",
            "  inflating: data/train/6/6381.png   \n",
            "  inflating: data/train/6/6385.png   \n",
            "  inflating: data/train/6/6388.png   \n",
            "  inflating: data/train/6/6395.png   \n",
            "  inflating: data/train/6/6397.png   \n",
            "  inflating: data/train/6/6398.png   \n",
            "  inflating: data/train/6/6402.png   \n",
            "  inflating: data/train/6/6405.png   \n",
            "  inflating: data/train/6/6412.png   \n",
            "  inflating: data/train/6/6423.png   \n",
            "  inflating: data/train/6/643.png    \n",
            "  inflating: data/train/6/644.png    \n",
            "  inflating: data/train/6/6443.png   \n",
            "  inflating: data/train/6/6446.png   \n",
            "  inflating: data/train/6/6453.png   \n",
            "  inflating: data/train/6/6456.png   \n",
            "  inflating: data/train/6/6463.png   \n",
            "  inflating: data/train/6/6476.png   \n",
            "  inflating: data/train/6/6480.png   \n",
            "  inflating: data/train/6/649.png    \n",
            "  inflating: data/train/6/6499.png   \n",
            "  inflating: data/train/6/6502.png   \n",
            "  inflating: data/train/6/6503.png   \n",
            "  inflating: data/train/6/651.png    \n",
            "  inflating: data/train/6/6523.png   \n",
            "  inflating: data/train/6/6527.png   \n",
            "  inflating: data/train/6/6528.png   \n",
            "  inflating: data/train/6/6540.png   \n",
            "  inflating: data/train/6/6545.png   \n",
            "  inflating: data/train/6/6552.png   \n",
            "  inflating: data/train/6/6554.png   \n",
            "  inflating: data/train/6/6582.png   \n",
            "  inflating: data/train/6/6585.png   \n",
            "  inflating: data/train/6/6587.png   \n",
            "  inflating: data/train/6/6590.png   \n",
            "  inflating: data/train/6/6593.png   \n",
            "  inflating: data/train/6/6596.png   \n",
            "  inflating: data/train/6/6599.png   \n",
            "  inflating: data/train/6/6602.png   \n",
            "  inflating: data/train/6/6613.png   \n",
            "  inflating: data/train/6/662.png    \n",
            "  inflating: data/train/6/6630.png   \n",
            "  inflating: data/train/6/6635.png   \n",
            "  inflating: data/train/6/6636.png   \n",
            "  inflating: data/train/6/6641.png   \n",
            "  inflating: data/train/6/6648.png   \n",
            "  inflating: data/train/6/6661.png   \n",
            "  inflating: data/train/6/6664.png   \n",
            "  inflating: data/train/6/6679.png   \n",
            "  inflating: data/train/6/6682.png   \n",
            "  inflating: data/train/6/6683.png   \n",
            "  inflating: data/train/6/6684.png   \n",
            "  inflating: data/train/6/6688.png   \n",
            "  inflating: data/train/6/6692.png   \n",
            "  inflating: data/train/6/6705.png   \n",
            "  inflating: data/train/6/6709.png   \n",
            "  inflating: data/train/6/6710.png   \n",
            "  inflating: data/train/6/6718.png   \n",
            "  inflating: data/train/6/672.png    \n",
            "  inflating: data/train/6/6727.png   \n",
            "  inflating: data/train/6/673.png    \n",
            "  inflating: data/train/6/6741.png   \n",
            "  inflating: data/train/6/6743.png   \n",
            "  inflating: data/train/6/6757.png   \n",
            "  inflating: data/train/6/6769.png   \n",
            "  inflating: data/train/6/6772.png   \n",
            "  inflating: data/train/6/6785.png   \n",
            "  inflating: data/train/6/679.png    \n",
            "  inflating: data/train/6/6790.png   \n",
            "  inflating: data/train/6/6800.png   \n",
            "  inflating: data/train/6/6805.png   \n",
            "  inflating: data/train/6/6806.png   \n",
            "  inflating: data/train/6/6810.png   \n",
            "  inflating: data/train/6/6813.png   \n",
            "  inflating: data/train/6/6825.png   \n",
            "  inflating: data/train/6/6826.png   \n",
            "  inflating: data/train/6/6836.png   \n",
            "  inflating: data/train/6/6838.png   \n",
            "  inflating: data/train/6/6842.png   \n",
            "  inflating: data/train/6/6857.png   \n",
            "  inflating: data/train/6/6859.png   \n",
            "  inflating: data/train/6/6862.png   \n",
            "  inflating: data/train/6/6869.png   \n",
            "  inflating: data/train/6/6873.png   \n",
            "  inflating: data/train/6/6888.png   \n",
            "  inflating: data/train/6/6891.png   \n",
            "  inflating: data/train/6/6892.png   \n",
            "  inflating: data/train/6/6903.png   \n",
            "  inflating: data/train/6/6911.png   \n",
            "  inflating: data/train/6/6920.png   \n",
            "  inflating: data/train/6/6924.png   \n",
            "  inflating: data/train/6/6925.png   \n",
            "  inflating: data/train/6/6932.png   \n",
            "  inflating: data/train/6/6934.png   \n",
            "  inflating: data/train/6/694.png    \n",
            "  inflating: data/train/6/6949.png   \n",
            "  inflating: data/train/6/6950.png   \n",
            "  inflating: data/train/6/6953.png   \n",
            "  inflating: data/train/6/6956.png   \n",
            "  inflating: data/train/6/6967.png   \n",
            "  inflating: data/train/6/6969.png   \n",
            "  inflating: data/train/6/6971.png   \n",
            "  inflating: data/train/6/6972.png   \n",
            "  inflating: data/train/6/6974.png   \n",
            "  inflating: data/train/6/6981.png   \n",
            "  inflating: data/train/6/6983.png   \n",
            "  inflating: data/train/6/699.png    \n",
            "  inflating: data/train/6/6998.png   \n",
            "  inflating: data/train/6/7008.png   \n",
            "  inflating: data/train/6/7009.png   \n",
            "  inflating: data/train/6/7041.png   \n",
            "  inflating: data/train/6/7043.png   \n",
            "  inflating: data/train/6/7047.png   \n",
            "  inflating: data/train/6/705.png    \n",
            "  inflating: data/train/6/7054.png   \n",
            "  inflating: data/train/6/7055.png   \n",
            "  inflating: data/train/6/7064.png   \n",
            "  inflating: data/train/6/7070.png   \n",
            "  inflating: data/train/6/7072.png   \n",
            "  inflating: data/train/6/7077.png   \n",
            "  inflating: data/train/6/7092.png   \n",
            "  inflating: data/train/6/7094.png   \n",
            "  inflating: data/train/6/7099.png   \n",
            "  inflating: data/train/6/7106.png   \n",
            "  inflating: data/train/6/7107.png   \n",
            "  inflating: data/train/6/7110.png   \n",
            "  inflating: data/train/6/7118.png   \n",
            "  inflating: data/train/6/7127.png   \n",
            "  inflating: data/train/6/7128.png   \n",
            "  inflating: data/train/6/7131.png   \n",
            "  inflating: data/train/6/7132.png   \n",
            "  inflating: data/train/6/7137.png   \n",
            "  inflating: data/train/6/7145.png   \n",
            "  inflating: data/train/6/7149.png   \n",
            "  inflating: data/train/6/715.png    \n",
            "  inflating: data/train/6/7155.png   \n",
            "  inflating: data/train/6/7156.png   \n",
            "  inflating: data/train/6/7169.png   \n",
            "  inflating: data/train/6/717.png    \n",
            "  inflating: data/train/6/7198.png   \n",
            "  inflating: data/train/6/7205.png   \n",
            "  inflating: data/train/6/7214.png   \n",
            "  inflating: data/train/6/7220.png   \n",
            "  inflating: data/train/6/7227.png   \n",
            "  inflating: data/train/6/7229.png   \n",
            "  inflating: data/train/6/7239.png   \n",
            "  inflating: data/train/6/7246.png   \n",
            "  inflating: data/train/6/725.png    \n",
            "  inflating: data/train/6/7254.png   \n",
            "  inflating: data/train/6/7258.png   \n",
            "  inflating: data/train/6/7273.png   \n",
            "  inflating: data/train/6/7274.png   \n",
            "  inflating: data/train/6/7276.png   \n",
            "  inflating: data/train/6/7278.png   \n",
            "  inflating: data/train/6/7294.png   \n",
            "  inflating: data/train/6/7297.png   \n",
            "  inflating: data/train/6/7304.png   \n",
            "  inflating: data/train/6/7312.png   \n",
            "  inflating: data/train/6/7316.png   \n",
            "  inflating: data/train/6/7319.png   \n",
            "  inflating: data/train/6/732.png    \n",
            "  inflating: data/train/6/7324.png   \n",
            "  inflating: data/train/6/7326.png   \n",
            "  inflating: data/train/6/7344.png   \n",
            "  inflating: data/train/6/7348.png   \n",
            "  inflating: data/train/6/7361.png   \n",
            "  inflating: data/train/6/7368.png   \n",
            "  inflating: data/train/6/7376.png   \n",
            "  inflating: data/train/6/7391.png   \n",
            "  inflating: data/train/6/7395.png   \n",
            "  inflating: data/train/6/7397.png   \n",
            "  inflating: data/train/6/7398.png   \n",
            "  inflating: data/train/6/7399.png   \n",
            "  inflating: data/train/6/7401.png   \n",
            "  inflating: data/train/6/7402.png   \n",
            "  inflating: data/train/6/7403.png   \n",
            "  inflating: data/train/6/7404.png   \n",
            "  inflating: data/train/6/7408.png   \n",
            "  inflating: data/train/6/7414.png   \n",
            "  inflating: data/train/6/7422.png   \n",
            "  inflating: data/train/6/7423.png   \n",
            "  inflating: data/train/6/7433.png   \n",
            "  inflating: data/train/6/7435.png   \n",
            "  inflating: data/train/6/7443.png   \n",
            "  inflating: data/train/6/7447.png   \n",
            "  inflating: data/train/6/7448.png   \n",
            "  inflating: data/train/6/7456.png   \n",
            "  inflating: data/train/6/7462.png   \n",
            "  inflating: data/train/6/7466.png   \n",
            "  inflating: data/train/6/747.png    \n",
            "  inflating: data/train/6/7470.png   \n",
            "  inflating: data/train/6/7474.png   \n",
            "  inflating: data/train/6/7486.png   \n",
            "  inflating: data/train/6/7490.png   \n",
            "  inflating: data/train/6/7494.png   \n",
            "  inflating: data/train/6/7495.png   \n",
            "  inflating: data/train/6/7498.png   \n",
            "  inflating: data/train/6/75.png     \n",
            "  inflating: data/train/6/7507.png   \n",
            "  inflating: data/train/6/7515.png   \n",
            "  inflating: data/train/6/7516.png   \n",
            "  inflating: data/train/6/7519.png   \n",
            "  inflating: data/train/6/7525.png   \n",
            "  inflating: data/train/6/7528.png   \n",
            "  inflating: data/train/6/7530.png   \n",
            "  inflating: data/train/6/7538.png   \n",
            "  inflating: data/train/6/7551.png   \n",
            "  inflating: data/train/6/757.png    \n",
            "  inflating: data/train/6/7571.png   \n",
            "  inflating: data/train/6/7575.png   \n",
            "  inflating: data/train/6/7588.png   \n",
            "  inflating: data/train/6/7599.png   \n",
            "  inflating: data/train/6/7606.png   \n",
            "  inflating: data/train/6/7607.png   \n",
            "  inflating: data/train/6/7608.png   \n",
            "  inflating: data/train/6/7611.png   \n",
            "  inflating: data/train/6/762.png    \n",
            "  inflating: data/train/6/7624.png   \n",
            "  inflating: data/train/6/7628.png   \n",
            "  inflating: data/train/6/7630.png   \n",
            "  inflating: data/train/6/7631.png   \n",
            "  inflating: data/train/6/7643.png   \n",
            "  inflating: data/train/6/7644.png   \n",
            "  inflating: data/train/6/7650.png   \n",
            "  inflating: data/train/6/766.png    \n",
            "  inflating: data/train/6/7660.png   \n",
            "  inflating: data/train/6/7663.png   \n",
            "  inflating: data/train/6/7667.png   \n",
            "  inflating: data/train/6/767.png    \n",
            "  inflating: data/train/6/7686.png   \n",
            "  inflating: data/train/6/7694.png   \n",
            "  inflating: data/train/6/7695.png   \n",
            "  inflating: data/train/6/7696.png   \n",
            "  inflating: data/train/6/7701.png   \n",
            "  inflating: data/train/6/7709.png   \n",
            "  inflating: data/train/6/7717.png   \n",
            "  inflating: data/train/6/772.png    \n",
            "  inflating: data/train/6/7725.png   \n",
            "  inflating: data/train/6/7727.png   \n",
            "  inflating: data/train/6/7728.png   \n",
            "  inflating: data/train/6/7738.png   \n",
            "  inflating: data/train/6/7745.png   \n",
            "  inflating: data/train/6/7754.png   \n",
            "  inflating: data/train/6/7755.png   \n",
            "  inflating: data/train/6/7758.png   \n",
            "  inflating: data/train/6/7765.png   \n",
            "  inflating: data/train/6/7769.png   \n",
            "  inflating: data/train/6/7772.png   \n",
            "  inflating: data/train/6/7773.png   \n",
            "  inflating: data/train/6/7775.png   \n",
            "  inflating: data/train/6/7788.png   \n",
            "  inflating: data/train/6/7789.png   \n",
            "  inflating: data/train/6/7803.png   \n",
            "  inflating: data/train/6/781.png    \n",
            "  inflating: data/train/6/7812.png   \n",
            "  inflating: data/train/6/7814.png   \n",
            "  inflating: data/train/6/7823.png   \n",
            "  inflating: data/train/6/7831.png   \n",
            "  inflating: data/train/6/7834.png   \n",
            "  inflating: data/train/6/7842.png   \n",
            "  inflating: data/train/6/785.png    \n",
            "  inflating: data/train/6/7867.png   \n",
            "  inflating: data/train/6/787.png    \n",
            "  inflating: data/train/6/7871.png   \n",
            "  inflating: data/train/6/7882.png   \n",
            "  inflating: data/train/6/7884.png   \n",
            "  inflating: data/train/6/7890.png   \n",
            "  inflating: data/train/6/7903.png   \n",
            "  inflating: data/train/6/7909.png   \n",
            "  inflating: data/train/6/7922.png   \n",
            "  inflating: data/train/6/7927.png   \n",
            "  inflating: data/train/6/793.png    \n",
            "  inflating: data/train/6/7932.png   \n",
            "  inflating: data/train/6/7934.png   \n",
            "  inflating: data/train/6/7935.png   \n",
            "  inflating: data/train/6/7939.png   \n",
            "  inflating: data/train/6/7946.png   \n",
            "  inflating: data/train/6/7947.png   \n",
            "  inflating: data/train/6/795.png    \n",
            "  inflating: data/train/6/7953.png   \n",
            "  inflating: data/train/6/7954.png   \n",
            "  inflating: data/train/6/7958.png   \n",
            "  inflating: data/train/6/7968.png   \n",
            "  inflating: data/train/6/8003.png   \n",
            "  inflating: data/train/6/8004.png   \n",
            "  inflating: data/train/6/8007.png   \n",
            "  inflating: data/train/6/801.png    \n",
            "  inflating: data/train/6/8011.png   \n",
            "  inflating: data/train/6/8019.png   \n",
            "  inflating: data/train/6/8023.png   \n",
            "  inflating: data/train/6/8048.png   \n",
            "  inflating: data/train/6/8054.png   \n",
            "  inflating: data/train/6/8058.png   \n",
            "  inflating: data/train/6/8064.png   \n",
            "  inflating: data/train/6/8066.png   \n",
            "  inflating: data/train/6/8068.png   \n",
            "  inflating: data/train/6/8071.png   \n",
            "  inflating: data/train/6/8073.png   \n",
            "  inflating: data/train/6/8075.png   \n",
            "  inflating: data/train/6/8076.png   \n",
            "  inflating: data/train/6/8083.png   \n",
            "  inflating: data/train/6/8090.png   \n",
            "  inflating: data/train/6/8100.png   \n",
            "  inflating: data/train/6/8105.png   \n",
            "  inflating: data/train/6/8107.png   \n",
            "  inflating: data/train/6/8125.png   \n",
            "  inflating: data/train/6/8126.png   \n",
            "  inflating: data/train/6/8127.png   \n",
            "  inflating: data/train/6/8131.png   \n",
            "  inflating: data/train/6/8141.png   \n",
            "  inflating: data/train/6/8142.png   \n",
            "  inflating: data/train/6/8145.png   \n",
            "  inflating: data/train/6/8154.png   \n",
            "  inflating: data/train/6/8179.png   \n",
            "  inflating: data/train/6/8181.png   \n",
            "  inflating: data/train/6/8186.png   \n",
            "  inflating: data/train/6/8193.png   \n",
            "  inflating: data/train/6/8195.png   \n",
            "  inflating: data/train/6/8211.png   \n",
            "  inflating: data/train/6/8213.png   \n",
            "  inflating: data/train/6/8214.png   \n",
            "  inflating: data/train/6/8216.png   \n",
            "  inflating: data/train/6/8230.png   \n",
            "  inflating: data/train/6/8233.png   \n",
            "  inflating: data/train/6/8234.png   \n",
            "  inflating: data/train/6/8241.png   \n",
            "  inflating: data/train/6/8247.png   \n",
            "  inflating: data/train/6/8249.png   \n",
            "  inflating: data/train/6/8262.png   \n",
            "  inflating: data/train/6/8264.png   \n",
            "  inflating: data/train/6/8277.png   \n",
            "  inflating: data/train/6/8280.png   \n",
            "  inflating: data/train/6/8283.png   \n",
            "  inflating: data/train/6/8285.png   \n",
            "  inflating: data/train/6/8295.png   \n",
            "  inflating: data/train/6/83.png     \n",
            "  inflating: data/train/6/8308.png   \n",
            "  inflating: data/train/6/8311.png   \n",
            "  inflating: data/train/6/8312.png   \n",
            "  inflating: data/train/6/8314.png   \n",
            "  inflating: data/train/6/8318.png   \n",
            "  inflating: data/train/6/8319.png   \n",
            "  inflating: data/train/6/832.png    \n",
            "  inflating: data/train/6/8341.png   \n",
            "  inflating: data/train/6/8347.png   \n",
            "  inflating: data/train/6/8352.png   \n",
            "  inflating: data/train/6/8357.png   \n",
            "  inflating: data/train/6/8362.png   \n",
            "  inflating: data/train/6/8365.png   \n",
            "  inflating: data/train/6/8373.png   \n",
            "  inflating: data/train/6/8375.png   \n",
            "  inflating: data/train/6/838.png    \n",
            "  inflating: data/train/6/8381.png   \n",
            "  inflating: data/train/6/8384.png   \n",
            "  inflating: data/train/6/8386.png   \n",
            "  inflating: data/train/6/8388.png   \n",
            "  inflating: data/train/6/8389.png   \n",
            "  inflating: data/train/6/8392.png   \n",
            "  inflating: data/train/6/8394.png   \n",
            "  inflating: data/train/6/8400.png   \n",
            "  inflating: data/train/6/8407.png   \n",
            "  inflating: data/train/6/841.png    \n",
            "  inflating: data/train/6/8410.png   \n",
            "  inflating: data/train/6/8411.png   \n",
            "  inflating: data/train/6/8414.png   \n",
            "  inflating: data/train/6/8425.png   \n",
            "  inflating: data/train/6/8428.png   \n",
            "  inflating: data/train/6/8441.png   \n",
            "  inflating: data/train/6/8443.png   \n",
            "  inflating: data/train/6/8450.png   \n",
            "  inflating: data/train/6/8451.png   \n",
            "  inflating: data/train/6/8471.png   \n",
            "  inflating: data/train/6/8483.png   \n",
            "  inflating: data/train/6/8491.png   \n",
            "  inflating: data/train/6/8497.png   \n",
            "  inflating: data/train/6/8498.png   \n",
            "  inflating: data/train/6/8514.png   \n",
            "  inflating: data/train/6/8516.png   \n",
            "  inflating: data/train/6/852.png    \n",
            "  inflating: data/train/6/8527.png   \n",
            "  inflating: data/train/6/8533.png   \n",
            "  inflating: data/train/6/8536.png   \n",
            "  inflating: data/train/6/8538.png   \n",
            "  inflating: data/train/6/8546.png   \n",
            "  inflating: data/train/6/855.png    \n",
            "  inflating: data/train/6/8558.png   \n",
            "  inflating: data/train/6/8567.png   \n",
            "  inflating: data/train/6/8570.png   \n",
            "  inflating: data/train/6/8575.png   \n",
            "  inflating: data/train/6/858.png    \n",
            "  inflating: data/train/6/8582.png   \n",
            "  inflating: data/train/6/8587.png   \n",
            "  inflating: data/train/6/8597.png   \n",
            "  inflating: data/train/6/8599.png   \n",
            "  inflating: data/train/6/8601.png   \n",
            "  inflating: data/train/6/8603.png   \n",
            "  inflating: data/train/6/8610.png   \n",
            "  inflating: data/train/6/8615.png   \n",
            "  inflating: data/train/6/8617.png   \n",
            "  inflating: data/train/6/8635.png   \n",
            "  inflating: data/train/6/8642.png   \n",
            "  inflating: data/train/6/8643.png   \n",
            "  inflating: data/train/6/8646.png   \n",
            "  inflating: data/train/6/8648.png   \n",
            "  inflating: data/train/6/8663.png   \n",
            "  inflating: data/train/6/8664.png   \n",
            "  inflating: data/train/6/8666.png   \n",
            "  inflating: data/train/6/8670.png   \n",
            "  inflating: data/train/6/8680.png   \n",
            "  inflating: data/train/6/8687.png   \n",
            "  inflating: data/train/6/8690.png   \n",
            "  inflating: data/train/6/8694.png   \n",
            "  inflating: data/train/6/8697.png   \n",
            "  inflating: data/train/6/8707.png   \n",
            "  inflating: data/train/6/8711.png   \n",
            "  inflating: data/train/6/8724.png   \n",
            "  inflating: data/train/6/8732.png   \n",
            "  inflating: data/train/6/8739.png   \n",
            "  inflating: data/train/6/874.png    \n",
            "  inflating: data/train/6/8748.png   \n",
            "  inflating: data/train/6/875.png    \n",
            "  inflating: data/train/6/8750.png   \n",
            "  inflating: data/train/6/8752.png   \n",
            "  inflating: data/train/6/8755.png   \n",
            "  inflating: data/train/6/8757.png   \n",
            "  inflating: data/train/6/8763.png   \n",
            "  inflating: data/train/6/8765.png   \n",
            "  inflating: data/train/6/8768.png   \n",
            "  inflating: data/train/6/877.png    \n",
            "  inflating: data/train/6/8770.png   \n",
            "  inflating: data/train/6/8782.png   \n",
            "  inflating: data/train/6/8784.png   \n",
            "  inflating: data/train/6/8785.png   \n",
            "  inflating: data/train/6/8788.png   \n",
            "  inflating: data/train/6/8791.png   \n",
            "  inflating: data/train/6/8798.png   \n",
            "  inflating: data/train/6/8799.png   \n",
            "  inflating: data/train/6/8802.png   \n",
            "  inflating: data/train/6/8807.png   \n",
            "  inflating: data/train/6/8808.png   \n",
            "  inflating: data/train/6/8809.png   \n",
            "  inflating: data/train/6/8815.png   \n",
            "  inflating: data/train/6/8817.png   \n",
            "  inflating: data/train/6/8829.png   \n",
            "  inflating: data/train/6/8832.png   \n",
            "  inflating: data/train/6/884.png    \n",
            "  inflating: data/train/6/8842.png   \n",
            "  inflating: data/train/6/8843.png   \n",
            "  inflating: data/train/6/885.png    \n",
            "  inflating: data/train/6/8850.png   \n",
            "  inflating: data/train/6/8854.png   \n",
            "  inflating: data/train/6/8860.png   \n",
            "  inflating: data/train/6/8862.png   \n",
            "  inflating: data/train/6/8863.png   \n",
            "  inflating: data/train/6/8868.png   \n",
            "  inflating: data/train/6/8872.png   \n",
            "  inflating: data/train/6/8874.png   \n",
            "  inflating: data/train/6/8880.png   \n",
            "  inflating: data/train/6/8882.png   \n",
            "  inflating: data/train/6/8883.png   \n",
            "  inflating: data/train/6/8884.png   \n",
            "  inflating: data/train/6/8891.png   \n",
            "  inflating: data/train/6/8894.png   \n",
            "  inflating: data/train/6/8898.png   \n",
            "  inflating: data/train/6/8905.png   \n",
            "  inflating: data/train/6/8907.png   \n",
            "  inflating: data/train/6/891.png    \n",
            "  inflating: data/train/6/8910.png   \n",
            "  inflating: data/train/6/8912.png   \n",
            "  inflating: data/train/6/8919.png   \n",
            "  inflating: data/train/6/8927.png   \n",
            "  inflating: data/train/6/8937.png   \n",
            "  inflating: data/train/6/8946.png   \n",
            "  inflating: data/train/6/8951.png   \n",
            "  inflating: data/train/6/8968.png   \n",
            "  inflating: data/train/6/8972.png   \n",
            "  inflating: data/train/6/8976.png   \n",
            "  inflating: data/train/6/8980.png   \n",
            "  inflating: data/train/6/8984.png   \n",
            "  inflating: data/train/6/8986.png   \n",
            "  inflating: data/train/6/8987.png   \n",
            "  inflating: data/train/6/8991.png   \n",
            "  inflating: data/train/6/8995.png   \n",
            "  inflating: data/train/6/9000.png   \n",
            "  inflating: data/train/6/901.png    \n",
            "  inflating: data/train/6/9011.png   \n",
            "  inflating: data/train/6/9014.png   \n",
            "  inflating: data/train/6/9016.png   \n",
            "  inflating: data/train/6/9018.png   \n",
            "  inflating: data/train/6/9029.png   \n",
            "  inflating: data/train/6/9039.png   \n",
            "  inflating: data/train/6/9045.png   \n",
            "  inflating: data/train/6/9046.png   \n",
            "  inflating: data/train/6/905.png    \n",
            "  inflating: data/train/6/9050.png   \n",
            "  inflating: data/train/6/9052.png   \n",
            "  inflating: data/train/6/9054.png   \n",
            "  inflating: data/train/6/9055.png   \n",
            "  inflating: data/train/6/9061.png   \n",
            "  inflating: data/train/6/9064.png   \n",
            "  inflating: data/train/6/9069.png   \n",
            "  inflating: data/train/6/9072.png   \n",
            "  inflating: data/train/6/9074.png   \n",
            "  inflating: data/train/6/9076.png   \n",
            "  inflating: data/train/6/9080.png   \n",
            "  inflating: data/train/6/9083.png   \n",
            "  inflating: data/train/6/9092.png   \n",
            "  inflating: data/train/6/9099.png   \n",
            "  inflating: data/train/6/9106.png   \n",
            "  inflating: data/train/6/912.png    \n",
            "  inflating: data/train/6/9128.png   \n",
            "  inflating: data/train/6/913.png    \n",
            "  inflating: data/train/6/9143.png   \n",
            "  inflating: data/train/6/9149.png   \n",
            "  inflating: data/train/6/916.png    \n",
            "  inflating: data/train/6/917.png    \n",
            "  inflating: data/train/6/9170.png   \n",
            "  inflating: data/train/6/9175.png   \n",
            "  inflating: data/train/6/9183.png   \n",
            "  inflating: data/train/6/9186.png   \n",
            "  inflating: data/train/6/919.png    \n",
            "  inflating: data/train/6/9190.png   \n",
            "  inflating: data/train/6/9195.png   \n",
            "  inflating: data/train/6/92.png     \n",
            "  inflating: data/train/6/9209.png   \n",
            "  inflating: data/train/6/9210.png   \n",
            "  inflating: data/train/6/9215.png   \n",
            "  inflating: data/train/6/9218.png   \n",
            "  inflating: data/train/6/9221.png   \n",
            "  inflating: data/train/6/9223.png   \n",
            "  inflating: data/train/6/9224.png   \n",
            "  inflating: data/train/6/9229.png   \n",
            "  inflating: data/train/6/9230.png   \n",
            "  inflating: data/train/6/9235.png   \n",
            "  inflating: data/train/6/9236.png   \n",
            "  inflating: data/train/6/9243.png   \n",
            "  inflating: data/train/6/925.png    \n",
            "  inflating: data/train/6/9250.png   \n",
            "  inflating: data/train/6/9252.png   \n",
            "  inflating: data/train/6/9254.png   \n",
            "  inflating: data/train/6/9259.png   \n",
            "  inflating: data/train/6/9262.png   \n",
            "  inflating: data/train/6/9265.png   \n",
            "  inflating: data/train/6/9273.png   \n",
            "  inflating: data/train/6/9276.png   \n",
            "  inflating: data/train/6/9288.png   \n",
            "  inflating: data/train/6/9292.png   \n",
            "  inflating: data/train/6/9295.png   \n",
            "  inflating: data/train/6/930.png    \n",
            "  inflating: data/train/6/9303.png   \n",
            "  inflating: data/train/6/9304.png   \n",
            "  inflating: data/train/6/9305.png   \n",
            "  inflating: data/train/6/931.png    \n",
            "  inflating: data/train/6/9314.png   \n",
            "  inflating: data/train/6/9324.png   \n",
            "  inflating: data/train/6/9335.png   \n",
            "  inflating: data/train/6/9349.png   \n",
            "  inflating: data/train/6/9355.png   \n",
            "  inflating: data/train/6/9356.png   \n",
            "  inflating: data/train/6/936.png    \n",
            "  inflating: data/train/6/9360.png   \n",
            "  inflating: data/train/6/9362.png   \n",
            "  inflating: data/train/6/9363.png   \n",
            "  inflating: data/train/6/937.png    \n",
            "  inflating: data/train/6/9381.png   \n",
            "  inflating: data/train/6/9390.png   \n",
            "  inflating: data/train/6/9396.png   \n",
            "  inflating: data/train/6/94.png     \n",
            "  inflating: data/train/6/9409.png   \n",
            "  inflating: data/train/6/9410.png   \n",
            "  inflating: data/train/6/9414.png   \n",
            "  inflating: data/train/6/9415.png   \n",
            "  inflating: data/train/6/9420.png   \n",
            "  inflating: data/train/6/9422.png   \n",
            "  inflating: data/train/6/9427.png   \n",
            "  inflating: data/train/6/9430.png   \n",
            "  inflating: data/train/6/9434.png   \n",
            "  inflating: data/train/6/9438.png   \n",
            "  inflating: data/train/6/9439.png   \n",
            "  inflating: data/train/6/9441.png   \n",
            "  inflating: data/train/6/9444.png   \n",
            "  inflating: data/train/6/945.png    \n",
            "  inflating: data/train/6/9451.png   \n",
            "  inflating: data/train/6/9453.png   \n",
            "  inflating: data/train/6/9466.png   \n",
            "  inflating: data/train/6/9473.png   \n",
            "  inflating: data/train/6/9477.png   \n",
            "  inflating: data/train/6/9479.png   \n",
            "  inflating: data/train/6/9481.png   \n",
            "  inflating: data/train/6/9484.png   \n",
            "  inflating: data/train/6/9490.png   \n",
            "  inflating: data/train/6/9493.png   \n",
            "  inflating: data/train/6/9494.png   \n",
            "  inflating: data/train/6/9495.png   \n",
            "  inflating: data/train/6/9507.png   \n",
            "  inflating: data/train/6/9511.png   \n",
            "  inflating: data/train/6/952.png    \n",
            "  inflating: data/train/6/9522.png   \n",
            "  inflating: data/train/6/9529.png   \n",
            "  inflating: data/train/6/9543.png   \n",
            "  inflating: data/train/6/9550.png   \n",
            "  inflating: data/train/6/9558.png   \n",
            "  inflating: data/train/6/9559.png   \n",
            "  inflating: data/train/6/956.png    \n",
            "  inflating: data/train/6/9564.png   \n",
            "  inflating: data/train/6/958.png    \n",
            "  inflating: data/train/6/9580.png   \n",
            "  inflating: data/train/6/9581.png   \n",
            "  inflating: data/train/6/9584.png   \n",
            "  inflating: data/train/6/9589.png   \n",
            "  inflating: data/train/6/9601.png   \n",
            "  inflating: data/train/6/9604.png   \n",
            "  inflating: data/train/6/9609.png   \n",
            "  inflating: data/train/6/9611.png   \n",
            "  inflating: data/train/6/9627.png   \n",
            "  inflating: data/train/6/9633.png   \n",
            "  inflating: data/train/6/9638.png   \n",
            "  inflating: data/train/6/9666.png   \n",
            "  inflating: data/train/6/9669.png   \n",
            "  inflating: data/train/6/967.png    \n",
            "  inflating: data/train/6/9671.png   \n",
            "  inflating: data/train/6/9676.png   \n",
            "  inflating: data/train/6/9684.png   \n",
            "  inflating: data/train/6/9686.png   \n",
            "  inflating: data/train/6/9688.png   \n",
            "  inflating: data/train/6/9692.png   \n",
            "  inflating: data/train/6/970.png    \n",
            "  inflating: data/train/6/9700.png   \n",
            "  inflating: data/train/6/9701.png   \n",
            "  inflating: data/train/6/9710.png   \n",
            "  inflating: data/train/6/9711.png   \n",
            "  inflating: data/train/6/9712.png   \n",
            "  inflating: data/train/6/9717.png   \n",
            "  inflating: data/train/6/9718.png   \n",
            "  inflating: data/train/6/9722.png   \n",
            "  inflating: data/train/6/9723.png   \n",
            "  inflating: data/train/6/9724.png   \n",
            "  inflating: data/train/6/9736.png   \n",
            "  inflating: data/train/6/975.png    \n",
            "  inflating: data/train/6/9754.png   \n",
            "  inflating: data/train/6/9757.png   \n",
            "  inflating: data/train/6/9768.png   \n",
            "  inflating: data/train/6/9773.png   \n",
            "  inflating: data/train/6/9777.png   \n",
            "  inflating: data/train/6/9780.png   \n",
            "  inflating: data/train/6/979.png    \n",
            "  inflating: data/train/6/9793.png   \n",
            "  inflating: data/train/6/9799.png   \n",
            "  inflating: data/train/6/9803.png   \n",
            "  inflating: data/train/6/9805.png   \n",
            "  inflating: data/train/6/9807.png   \n",
            "  inflating: data/train/6/9823.png   \n",
            "  inflating: data/train/6/9824.png   \n",
            "  inflating: data/train/6/9829.png   \n",
            "  inflating: data/train/6/9831.png   \n",
            "  inflating: data/train/6/9838.png   \n",
            "  inflating: data/train/6/9843.png   \n",
            "  inflating: data/train/6/9852.png   \n",
            "  inflating: data/train/6/9858.png   \n",
            "  inflating: data/train/6/9867.png   \n",
            "  inflating: data/train/6/9871.png   \n",
            "  inflating: data/train/6/9874.png   \n",
            "  inflating: data/train/6/9881.png   \n",
            "  inflating: data/train/6/989.png    \n",
            "  inflating: data/train/6/9906.png   \n",
            "  inflating: data/train/6/9907.png   \n",
            "  inflating: data/train/6/991.png    \n",
            "  inflating: data/train/6/9915.png   \n",
            "  inflating: data/train/6/992.png    \n",
            "  inflating: data/train/6/9920.png   \n",
            "  inflating: data/train/6/9926.png   \n",
            "  inflating: data/train/6/9936.png   \n",
            "  inflating: data/train/6/9945.png   \n",
            "  inflating: data/train/6/9954.png   \n",
            "  inflating: data/train/6/996.png    \n",
            "  inflating: data/train/6/9970.png   \n",
            "  inflating: data/train/6/9975.png   \n",
            "  inflating: data/train/6/9980.png   \n",
            "  inflating: data/train/6/9982.png   \n",
            "  inflating: data/train/6/9987.png   \n",
            "  inflating: data/train/6/999.png    \n",
            "  inflating: data/train/6/9993.png   \n",
            "  inflating: data/train/6/9995.png   \n",
            "  inflating: data/train/6/9998.png   \n",
            "  inflating: data/train/6/9999.png   \n",
            "  inflating: data/val/0/28709.png    \n",
            "  inflating: data/val/0/28716.png    \n",
            "  inflating: data/val/0/28718.png    \n",
            "  inflating: data/val/0/28720.png    \n",
            "  inflating: data/val/0/28726.png    \n",
            "  inflating: data/val/0/28731.png    \n",
            "  inflating: data/val/0/28733.png    \n",
            "  inflating: data/val/0/28737.png    \n",
            "  inflating: data/val/0/28743.png    \n",
            "  inflating: data/val/0/28749.png    \n",
            "  inflating: data/val/0/28759.png    \n",
            "  inflating: data/val/0/28779.png    \n",
            "  inflating: data/val/0/28789.png    \n",
            "  inflating: data/val/0/28790.png    \n",
            "  inflating: data/val/0/28800.png    \n",
            "  inflating: data/val/0/28808.png    \n",
            "  inflating: data/val/0/28811.png    \n",
            "  inflating: data/val/0/28820.png    \n",
            "  inflating: data/val/0/28821.png    \n",
            "  inflating: data/val/0/28830.png    \n",
            "  inflating: data/val/0/28831.png    \n",
            "  inflating: data/val/0/28840.png    \n",
            "  inflating: data/val/0/28860.png    \n",
            "  inflating: data/val/0/28867.png    \n",
            "  inflating: data/val/0/28869.png    \n",
            "  inflating: data/val/0/28872.png    \n",
            "  inflating: data/val/0/28883.png    \n",
            "  inflating: data/val/0/28884.png    \n",
            "  inflating: data/val/0/28885.png    \n",
            "  inflating: data/val/0/28894.png    \n",
            "  inflating: data/val/0/28907.png    \n",
            "  inflating: data/val/0/28915.png    \n",
            "  inflating: data/val/0/28920.png    \n",
            "  inflating: data/val/0/28927.png    \n",
            "  inflating: data/val/0/28932.png    \n",
            "  inflating: data/val/0/28933.png    \n",
            "  inflating: data/val/0/28946.png    \n",
            "  inflating: data/val/0/28947.png    \n",
            "  inflating: data/val/0/28952.png    \n",
            "  inflating: data/val/0/28957.png    \n",
            "  inflating: data/val/0/28969.png    \n",
            "  inflating: data/val/0/28971.png    \n",
            "  inflating: data/val/0/28974.png    \n",
            "  inflating: data/val/0/28975.png    \n",
            "  inflating: data/val/0/28994.png    \n",
            "  inflating: data/val/0/28995.png    \n",
            "  inflating: data/val/0/28996.png    \n",
            "  inflating: data/val/0/29013.png    \n",
            "  inflating: data/val/0/29024.png    \n",
            "  inflating: data/val/0/29033.png    \n",
            "  inflating: data/val/0/29035.png    \n",
            "  inflating: data/val/0/29041.png    \n",
            "  inflating: data/val/0/29047.png    \n",
            "  inflating: data/val/0/29053.png    \n",
            "  inflating: data/val/0/29060.png    \n",
            "  inflating: data/val/0/29066.png    \n",
            "  inflating: data/val/0/29085.png    \n",
            "  inflating: data/val/0/29089.png    \n",
            "  inflating: data/val/0/29103.png    \n",
            "  inflating: data/val/0/29108.png    \n",
            "  inflating: data/val/0/29109.png    \n",
            "  inflating: data/val/0/29119.png    \n",
            "  inflating: data/val/0/29132.png    \n",
            "  inflating: data/val/0/29134.png    \n",
            "  inflating: data/val/0/29148.png    \n",
            "  inflating: data/val/0/29149.png    \n",
            "  inflating: data/val/0/29153.png    \n",
            "  inflating: data/val/0/29155.png    \n",
            "  inflating: data/val/0/29159.png    \n",
            "  inflating: data/val/0/29167.png    \n",
            "  inflating: data/val/0/29171.png    \n",
            "  inflating: data/val/0/29189.png    \n",
            "  inflating: data/val/0/29192.png    \n",
            "  inflating: data/val/0/29202.png    \n",
            "  inflating: data/val/0/29213.png    \n",
            "  inflating: data/val/0/29217.png    \n",
            "  inflating: data/val/0/29221.png    \n",
            "  inflating: data/val/0/29227.png    \n",
            "  inflating: data/val/0/29239.png    \n",
            "  inflating: data/val/0/29254.png    \n",
            "  inflating: data/val/0/29259.png    \n",
            "  inflating: data/val/0/29263.png    \n",
            "  inflating: data/val/0/29266.png    \n",
            "  inflating: data/val/0/29267.png    \n",
            "  inflating: data/val/0/29272.png    \n",
            "  inflating: data/val/0/29278.png    \n",
            "  inflating: data/val/0/29279.png    \n",
            "  inflating: data/val/0/29282.png    \n",
            "  inflating: data/val/0/29308.png    \n",
            "  inflating: data/val/0/29319.png    \n",
            "  inflating: data/val/0/29325.png    \n",
            "  inflating: data/val/0/29341.png    \n",
            "  inflating: data/val/0/29344.png    \n",
            "  inflating: data/val/0/29347.png    \n",
            "  inflating: data/val/0/29348.png    \n",
            "  inflating: data/val/0/29378.png    \n",
            "  inflating: data/val/0/29385.png    \n",
            "  inflating: data/val/0/29408.png    \n",
            "  inflating: data/val/0/29414.png    \n",
            "  inflating: data/val/0/29416.png    \n",
            "  inflating: data/val/0/29419.png    \n",
            "  inflating: data/val/0/29432.png    \n",
            "  inflating: data/val/0/29434.png    \n",
            "  inflating: data/val/0/29435.png    \n",
            "  inflating: data/val/0/29450.png    \n",
            "  inflating: data/val/0/29451.png    \n",
            "  inflating: data/val/0/29455.png    \n",
            "  inflating: data/val/0/29461.png    \n",
            "  inflating: data/val/0/29465.png    \n",
            "  inflating: data/val/0/29470.png    \n",
            "  inflating: data/val/0/29471.png    \n",
            "  inflating: data/val/0/29477.png    \n",
            "  inflating: data/val/0/29480.png    \n",
            "  inflating: data/val/0/29493.png    \n",
            "  inflating: data/val/0/29499.png    \n",
            "  inflating: data/val/0/29503.png    \n",
            "  inflating: data/val/0/29507.png    \n",
            "  inflating: data/val/0/29520.png    \n",
            "  inflating: data/val/0/29523.png    \n",
            "  inflating: data/val/0/29527.png    \n",
            "  inflating: data/val/0/29539.png    \n",
            "  inflating: data/val/0/29544.png    \n",
            "  inflating: data/val/0/29548.png    \n",
            "  inflating: data/val/0/29554.png    \n",
            "  inflating: data/val/0/29571.png    \n",
            "  inflating: data/val/0/29572.png    \n",
            "  inflating: data/val/0/29574.png    \n",
            "  inflating: data/val/0/29581.png    \n",
            "  inflating: data/val/0/29582.png    \n",
            "  inflating: data/val/0/29588.png    \n",
            "  inflating: data/val/0/29596.png    \n",
            "  inflating: data/val/0/29599.png    \n",
            "  inflating: data/val/0/29604.png    \n",
            "  inflating: data/val/0/29605.png    \n",
            "  inflating: data/val/0/29614.png    \n",
            "  inflating: data/val/0/29640.png    \n",
            "  inflating: data/val/0/29653.png    \n",
            "  inflating: data/val/0/29661.png    \n",
            "  inflating: data/val/0/29669.png    \n",
            "  inflating: data/val/0/29671.png    \n",
            "  inflating: data/val/0/29675.png    \n",
            "  inflating: data/val/0/29682.png    \n",
            "  inflating: data/val/0/29685.png    \n",
            "  inflating: data/val/0/29706.png    \n",
            "  inflating: data/val/0/29719.png    \n",
            "  inflating: data/val/0/29730.png    \n",
            "  inflating: data/val/0/29734.png    \n",
            "  inflating: data/val/0/29746.png    \n",
            "  inflating: data/val/0/29785.png    \n",
            "  inflating: data/val/0/29788.png    \n",
            "  inflating: data/val/0/29799.png    \n",
            "  inflating: data/val/0/29807.png    \n",
            "  inflating: data/val/0/29810.png    \n",
            "  inflating: data/val/0/29815.png    \n",
            "  inflating: data/val/0/29817.png    \n",
            "  inflating: data/val/0/29818.png    \n",
            "  inflating: data/val/0/29830.png    \n",
            "  inflating: data/val/0/29838.png    \n",
            "  inflating: data/val/0/29850.png    \n",
            "  inflating: data/val/0/29851.png    \n",
            "  inflating: data/val/0/29855.png    \n",
            "  inflating: data/val/0/29860.png    \n",
            "  inflating: data/val/0/29862.png    \n",
            "  inflating: data/val/0/29866.png    \n",
            "  inflating: data/val/0/29868.png    \n",
            "  inflating: data/val/0/29878.png    \n",
            "  inflating: data/val/0/29901.png    \n",
            "  inflating: data/val/0/29929.png    \n",
            "  inflating: data/val/0/29931.png    \n",
            "  inflating: data/val/0/29933.png    \n",
            "  inflating: data/val/0/29947.png    \n",
            "  inflating: data/val/0/29950.png    \n",
            "  inflating: data/val/0/29962.png    \n",
            "  inflating: data/val/0/29963.png    \n",
            "  inflating: data/val/0/29971.png    \n",
            "  inflating: data/val/0/29976.png    \n",
            "  inflating: data/val/0/29977.png    \n",
            "  inflating: data/val/0/29980.png    \n",
            "  inflating: data/val/0/30002.png    \n",
            "  inflating: data/val/0/30004.png    \n",
            "  inflating: data/val/0/30009.png    \n",
            "  inflating: data/val/0/30015.png    \n",
            "  inflating: data/val/0/30024.png    \n",
            "  inflating: data/val/0/30026.png    \n",
            "  inflating: data/val/0/30028.png    \n",
            "  inflating: data/val/0/30047.png    \n",
            "  inflating: data/val/0/30058.png    \n",
            "  inflating: data/val/0/30059.png    \n",
            "  inflating: data/val/0/30066.png    \n",
            "  inflating: data/val/0/30079.png    \n",
            "  inflating: data/val/0/30082.png    \n",
            "  inflating: data/val/0/30085.png    \n",
            "  inflating: data/val/0/30092.png    \n",
            "  inflating: data/val/0/30093.png    \n",
            "  inflating: data/val/0/30120.png    \n",
            "  inflating: data/val/0/30128.png    \n",
            "  inflating: data/val/0/30132.png    \n",
            "  inflating: data/val/0/30153.png    \n",
            "  inflating: data/val/0/30165.png    \n",
            "  inflating: data/val/0/30171.png    \n",
            "  inflating: data/val/0/30194.png    \n",
            "  inflating: data/val/0/30213.png    \n",
            "  inflating: data/val/0/30227.png    \n",
            "  inflating: data/val/0/30230.png    \n",
            "  inflating: data/val/0/30235.png    \n",
            "  inflating: data/val/0/30246.png    \n",
            "  inflating: data/val/0/30259.png    \n",
            "  inflating: data/val/0/30260.png    \n",
            "  inflating: data/val/0/30262.png    \n",
            "  inflating: data/val/0/30268.png    \n",
            "  inflating: data/val/0/30273.png    \n",
            "  inflating: data/val/0/30285.png    \n",
            "  inflating: data/val/0/30290.png    \n",
            "  inflating: data/val/0/30302.png    \n",
            "  inflating: data/val/0/30305.png    \n",
            "  inflating: data/val/0/30309.png    \n",
            "  inflating: data/val/0/30317.png    \n",
            "  inflating: data/val/0/30320.png    \n",
            "  inflating: data/val/0/30325.png    \n",
            "  inflating: data/val/0/30327.png    \n",
            "  inflating: data/val/0/30340.png    \n",
            "  inflating: data/val/0/30382.png    \n",
            "  inflating: data/val/0/30384.png    \n",
            "  inflating: data/val/0/30394.png    \n",
            "  inflating: data/val/0/30398.png    \n",
            "  inflating: data/val/0/30399.png    \n",
            "  inflating: data/val/0/30413.png    \n",
            "  inflating: data/val/0/30415.png    \n",
            "  inflating: data/val/0/30420.png    \n",
            "  inflating: data/val/0/30430.png    \n",
            "  inflating: data/val/0/30445.png    \n",
            "  inflating: data/val/0/30446.png    \n",
            "  inflating: data/val/0/30447.png    \n",
            "  inflating: data/val/0/30456.png    \n",
            "  inflating: data/val/0/30458.png    \n",
            "  inflating: data/val/0/30460.png    \n",
            "  inflating: data/val/0/30470.png    \n",
            "  inflating: data/val/0/30479.png    \n",
            "  inflating: data/val/0/30484.png    \n",
            "  inflating: data/val/0/30486.png    \n",
            "  inflating: data/val/0/30490.png    \n",
            "  inflating: data/val/0/30493.png    \n",
            "  inflating: data/val/0/30497.png    \n",
            "  inflating: data/val/0/30503.png    \n",
            "  inflating: data/val/0/30506.png    \n",
            "  inflating: data/val/0/30508.png    \n",
            "  inflating: data/val/0/30509.png    \n",
            "  inflating: data/val/0/30514.png    \n",
            "  inflating: data/val/0/30526.png    \n",
            "  inflating: data/val/0/30535.png    \n",
            "  inflating: data/val/0/30544.png    \n",
            "  inflating: data/val/0/30545.png    \n",
            "  inflating: data/val/0/30548.png    \n",
            "  inflating: data/val/0/30563.png    \n",
            "  inflating: data/val/0/30584.png    \n",
            "  inflating: data/val/0/30589.png    \n",
            "  inflating: data/val/0/30592.png    \n",
            "  inflating: data/val/0/30593.png    \n",
            "  inflating: data/val/0/30599.png    \n",
            "  inflating: data/val/0/30605.png    \n",
            "  inflating: data/val/0/30609.png    \n",
            "  inflating: data/val/0/30619.png    \n",
            "  inflating: data/val/0/30627.png    \n",
            "  inflating: data/val/0/30631.png    \n",
            "  inflating: data/val/0/30649.png    \n",
            "  inflating: data/val/0/30673.png    \n",
            "  inflating: data/val/0/30685.png    \n",
            "  inflating: data/val/0/30690.png    \n",
            "  inflating: data/val/0/30694.png    \n",
            "  inflating: data/val/0/30696.png    \n",
            "  inflating: data/val/0/30709.png    \n",
            "  inflating: data/val/0/30742.png    \n",
            "  inflating: data/val/0/30745.png    \n",
            "  inflating: data/val/0/30746.png    \n",
            "  inflating: data/val/0/30756.png    \n",
            "  inflating: data/val/0/30759.png    \n",
            "  inflating: data/val/0/30783.png    \n",
            "  inflating: data/val/0/30797.png    \n",
            "  inflating: data/val/0/30815.png    \n",
            "  inflating: data/val/0/30831.png    \n",
            "  inflating: data/val/0/30834.png    \n",
            "  inflating: data/val/0/30845.png    \n",
            "  inflating: data/val/0/30847.png    \n",
            "  inflating: data/val/0/30852.png    \n",
            "  inflating: data/val/0/30872.png    \n",
            "  inflating: data/val/0/30886.png    \n",
            "  inflating: data/val/0/30887.png    \n",
            "  inflating: data/val/0/30890.png    \n",
            "  inflating: data/val/0/30901.png    \n",
            "  inflating: data/val/0/30905.png    \n",
            "  inflating: data/val/0/30930.png    \n",
            "  inflating: data/val/0/30931.png    \n",
            "  inflating: data/val/0/30933.png    \n",
            "  inflating: data/val/0/30944.png    \n",
            "  inflating: data/val/0/30952.png    \n",
            "  inflating: data/val/0/30967.png    \n",
            "  inflating: data/val/0/30972.png    \n",
            "  inflating: data/val/0/30977.png    \n",
            "  inflating: data/val/0/30982.png    \n",
            "  inflating: data/val/0/31013.png    \n",
            "  inflating: data/val/0/31019.png    \n",
            "  inflating: data/val/0/31028.png    \n",
            "  inflating: data/val/0/31031.png    \n",
            "  inflating: data/val/0/31033.png    \n",
            "  inflating: data/val/0/31042.png    \n",
            "  inflating: data/val/0/31044.png    \n",
            "  inflating: data/val/0/31045.png    \n",
            "  inflating: data/val/0/31064.png    \n",
            "  inflating: data/val/0/31069.png    \n",
            "  inflating: data/val/0/31077.png    \n",
            "  inflating: data/val/0/31081.png    \n",
            "  inflating: data/val/0/31083.png    \n",
            "  inflating: data/val/0/31088.png    \n",
            "  inflating: data/val/0/31098.png    \n",
            "  inflating: data/val/0/31100.png    \n",
            "  inflating: data/val/0/31103.png    \n",
            "  inflating: data/val/0/31108.png    \n",
            "  inflating: data/val/0/31120.png    \n",
            "  inflating: data/val/0/31125.png    \n",
            "  inflating: data/val/0/31138.png    \n",
            "  inflating: data/val/0/31142.png    \n",
            "  inflating: data/val/0/31151.png    \n",
            "  inflating: data/val/0/31152.png    \n",
            "  inflating: data/val/0/31153.png    \n",
            "  inflating: data/val/0/31154.png    \n",
            "  inflating: data/val/0/31162.png    \n",
            "  inflating: data/val/0/31193.png    \n",
            "  inflating: data/val/0/31195.png    \n",
            "  inflating: data/val/0/31203.png    \n",
            "  inflating: data/val/0/31207.png    \n",
            "  inflating: data/val/0/31210.png    \n",
            "  inflating: data/val/0/31213.png    \n",
            "  inflating: data/val/0/31221.png    \n",
            "  inflating: data/val/0/31222.png    \n",
            "  inflating: data/val/0/31229.png    \n",
            "  inflating: data/val/0/31231.png    \n",
            "  inflating: data/val/0/31234.png    \n",
            "  inflating: data/val/0/31235.png    \n",
            "  inflating: data/val/0/31247.png    \n",
            "  inflating: data/val/0/31253.png    \n",
            "  inflating: data/val/0/31261.png    \n",
            "  inflating: data/val/0/31266.png    \n",
            "  inflating: data/val/0/31271.png    \n",
            "  inflating: data/val/0/31279.png    \n",
            "  inflating: data/val/0/31280.png    \n",
            "  inflating: data/val/0/31293.png    \n",
            "  inflating: data/val/0/31296.png    \n",
            "  inflating: data/val/0/31310.png    \n",
            "  inflating: data/val/0/31331.png    \n",
            "  inflating: data/val/0/31337.png    \n",
            "  inflating: data/val/0/31338.png    \n",
            "  inflating: data/val/0/31343.png    \n",
            "  inflating: data/val/0/31378.png    \n",
            "  inflating: data/val/0/31383.png    \n",
            "  inflating: data/val/0/31388.png    \n",
            "  inflating: data/val/0/31429.png    \n",
            "  inflating: data/val/0/31430.png    \n",
            "  inflating: data/val/0/31442.png    \n",
            "  inflating: data/val/0/31444.png    \n",
            "  inflating: data/val/0/31445.png    \n",
            "  inflating: data/val/0/31447.png    \n",
            "  inflating: data/val/0/31450.png    \n",
            "  inflating: data/val/0/31472.png    \n",
            "  inflating: data/val/0/31476.png    \n",
            "  inflating: data/val/0/31485.png    \n",
            "  inflating: data/val/0/31491.png    \n",
            "  inflating: data/val/0/31494.png    \n",
            "  inflating: data/val/0/31497.png    \n",
            "  inflating: data/val/0/31500.png    \n",
            "  inflating: data/val/0/31505.png    \n",
            "  inflating: data/val/0/31518.png    \n",
            "  inflating: data/val/0/31519.png    \n",
            "  inflating: data/val/0/31529.png    \n",
            "  inflating: data/val/0/31530.png    \n",
            "  inflating: data/val/0/31538.png    \n",
            "  inflating: data/val/0/31543.png    \n",
            "  inflating: data/val/0/31544.png    \n",
            "  inflating: data/val/0/31549.png    \n",
            "  inflating: data/val/0/31564.png    \n",
            "  inflating: data/val/0/31581.png    \n",
            "  inflating: data/val/0/31586.png    \n",
            "  inflating: data/val/0/31589.png    \n",
            "  inflating: data/val/0/31590.png    \n",
            "  inflating: data/val/0/31592.png    \n",
            "  inflating: data/val/0/31595.png    \n",
            "  inflating: data/val/0/31599.png    \n",
            "  inflating: data/val/0/31603.png    \n",
            "  inflating: data/val/0/31609.png    \n",
            "  inflating: data/val/0/31611.png    \n",
            "  inflating: data/val/0/31613.png    \n",
            "  inflating: data/val/0/31622.png    \n",
            "  inflating: data/val/0/31635.png    \n",
            "  inflating: data/val/0/31648.png    \n",
            "  inflating: data/val/0/31657.png    \n",
            "  inflating: data/val/0/31666.png    \n",
            "  inflating: data/val/0/31670.png    \n",
            "  inflating: data/val/0/31673.png    \n",
            "  inflating: data/val/0/31692.png    \n",
            "  inflating: data/val/0/31693.png    \n",
            "  inflating: data/val/0/31694.png    \n",
            "  inflating: data/val/0/31703.png    \n",
            "  inflating: data/val/0/31706.png    \n",
            "  inflating: data/val/0/31711.png    \n",
            "  inflating: data/val/0/31716.png    \n",
            "  inflating: data/val/0/31722.png    \n",
            "  inflating: data/val/0/31726.png    \n",
            "  inflating: data/val/0/31729.png    \n",
            "  inflating: data/val/0/31749.png    \n",
            "  inflating: data/val/0/31779.png    \n",
            "  inflating: data/val/0/31794.png    \n",
            "  inflating: data/val/0/31809.png    \n",
            "  inflating: data/val/0/31815.png    \n",
            "  inflating: data/val/0/31840.png    \n",
            "  inflating: data/val/0/31849.png    \n",
            "  inflating: data/val/0/31867.png    \n",
            "  inflating: data/val/0/31873.png    \n",
            "  inflating: data/val/0/31874.png    \n",
            "  inflating: data/val/0/31877.png    \n",
            "  inflating: data/val/0/31879.png    \n",
            "  inflating: data/val/0/31888.png    \n",
            "  inflating: data/val/0/31917.png    \n",
            "  inflating: data/val/0/31921.png    \n",
            "  inflating: data/val/0/31925.png    \n",
            "  inflating: data/val/0/31926.png    \n",
            "  inflating: data/val/0/31928.png    \n",
            "  inflating: data/val/0/31931.png    \n",
            "  inflating: data/val/0/31932.png    \n",
            "  inflating: data/val/0/31936.png    \n",
            "  inflating: data/val/0/31937.png    \n",
            "  inflating: data/val/0/31939.png    \n",
            "  inflating: data/val/0/31946.png    \n",
            "  inflating: data/val/0/31947.png    \n",
            "  inflating: data/val/0/31948.png    \n",
            "  inflating: data/val/0/31969.png    \n",
            "  inflating: data/val/0/31971.png    \n",
            "  inflating: data/val/0/32001.png    \n",
            "  inflating: data/val/0/32017.png    \n",
            "  inflating: data/val/0/32019.png    \n",
            "  inflating: data/val/0/32023.png    \n",
            "  inflating: data/val/0/32024.png    \n",
            "  inflating: data/val/0/32072.png    \n",
            "  inflating: data/val/0/32078.png    \n",
            "  inflating: data/val/0/32094.png    \n",
            "  inflating: data/val/0/32112.png    \n",
            "  inflating: data/val/0/32141.png    \n",
            "  inflating: data/val/0/32150.png    \n",
            "  inflating: data/val/0/32153.png    \n",
            "  inflating: data/val/0/32161.png    \n",
            "  inflating: data/val/0/32172.png    \n",
            "  inflating: data/val/0/32181.png    \n",
            "  inflating: data/val/0/32196.png    \n",
            "  inflating: data/val/0/32200.png    \n",
            "  inflating: data/val/0/32203.png    \n",
            "  inflating: data/val/0/32204.png    \n",
            "  inflating: data/val/0/32206.png    \n",
            "  inflating: data/val/0/32207.png    \n",
            "  inflating: data/val/0/32211.png    \n",
            "  inflating: data/val/0/32215.png    \n",
            "  inflating: data/val/0/32223.png    \n",
            "  inflating: data/val/0/32225.png    \n",
            "  inflating: data/val/0/32231.png    \n",
            "  inflating: data/val/0/32238.png    \n",
            "  inflating: data/val/0/32240.png    \n",
            "  inflating: data/val/0/32252.png    \n",
            "  inflating: data/val/0/32261.png    \n",
            "  inflating: data/val/0/32278.png    \n",
            "  inflating: data/val/0/32281.png    \n",
            "  inflating: data/val/1/28710.png    \n",
            "  inflating: data/val/1/28791.png    \n",
            "  inflating: data/val/1/28851.png    \n",
            "  inflating: data/val/1/28908.png    \n",
            "  inflating: data/val/1/28939.png    \n",
            "  inflating: data/val/1/28953.png    \n",
            "  inflating: data/val/1/28970.png    \n",
            "  inflating: data/val/1/28987.png    \n",
            "  inflating: data/val/1/29023.png    \n",
            "  inflating: data/val/1/29098.png    \n",
            "  inflating: data/val/1/29158.png    \n",
            "  inflating: data/val/1/29184.png    \n",
            "  inflating: data/val/1/29269.png    \n",
            "  inflating: data/val/1/29512.png    \n",
            "  inflating: data/val/1/29547.png    \n",
            "  inflating: data/val/1/29575.png    \n",
            "  inflating: data/val/1/29630.png    \n",
            "  inflating: data/val/1/29732.png    \n",
            "  inflating: data/val/1/29768.png    \n",
            "  inflating: data/val/1/29814.png    \n",
            "  inflating: data/val/1/29951.png    \n",
            "  inflating: data/val/1/30113.png    \n",
            "  inflating: data/val/1/30141.png    \n",
            "  inflating: data/val/1/30252.png    \n",
            "  inflating: data/val/1/30347.png    \n",
            "  inflating: data/val/1/30369.png    \n",
            "  inflating: data/val/1/30405.png    \n",
            "  inflating: data/val/1/30428.png    \n",
            "  inflating: data/val/1/30499.png    \n",
            "  inflating: data/val/1/30550.png    \n",
            "  inflating: data/val/1/30601.png    \n",
            "  inflating: data/val/1/30639.png    \n",
            "  inflating: data/val/1/30719.png    \n",
            "  inflating: data/val/1/30761.png    \n",
            "  inflating: data/val/1/30773.png    \n",
            "  inflating: data/val/1/30781.png    \n",
            "  inflating: data/val/1/30828.png    \n",
            "  inflating: data/val/1/30858.png    \n",
            "  inflating: data/val/1/30907.png    \n",
            "  inflating: data/val/1/30928.png    \n",
            "  inflating: data/val/1/31196.png    \n",
            "  inflating: data/val/1/31205.png    \n",
            "  inflating: data/val/1/31232.png    \n",
            "  inflating: data/val/1/31322.png    \n",
            "  inflating: data/val/1/31323.png    \n",
            "  inflating: data/val/1/31368.png    \n",
            "  inflating: data/val/1/31400.png    \n",
            "  inflating: data/val/1/31604.png    \n",
            "  inflating: data/val/1/31796.png    \n",
            "  inflating: data/val/1/31828.png    \n",
            "  inflating: data/val/1/32056.png    \n",
            "  inflating: data/val/1/32086.png    \n",
            "  inflating: data/val/1/32120.png    \n",
            "  inflating: data/val/1/32178.png    \n",
            "  inflating: data/val/1/32260.png    \n",
            "  inflating: data/val/1/32289.png    \n",
            "  inflating: data/val/2/28715.png    \n",
            "  inflating: data/val/2/28717.png    \n",
            "  inflating: data/val/2/28722.png    \n",
            "  inflating: data/val/2/28724.png    \n",
            "  inflating: data/val/2/28729.png    \n",
            "  inflating: data/val/2/28734.png    \n",
            "  inflating: data/val/2/28738.png    \n",
            "  inflating: data/val/2/28746.png    \n",
            "  inflating: data/val/2/28752.png    \n",
            "  inflating: data/val/2/28766.png    \n",
            "  inflating: data/val/2/28777.png    \n",
            "  inflating: data/val/2/28783.png    \n",
            "  inflating: data/val/2/28804.png    \n",
            "  inflating: data/val/2/28813.png    \n",
            "  inflating: data/val/2/28822.png    \n",
            "  inflating: data/val/2/28824.png    \n",
            "  inflating: data/val/2/28827.png    \n",
            "  inflating: data/val/2/28832.png    \n",
            "  inflating: data/val/2/28839.png    \n",
            "  inflating: data/val/2/28841.png    \n",
            "  inflating: data/val/2/28847.png    \n",
            "  inflating: data/val/2/28858.png    \n",
            "  inflating: data/val/2/28859.png    \n",
            "  inflating: data/val/2/28911.png    \n",
            "  inflating: data/val/2/28914.png    \n",
            "  inflating: data/val/2/28917.png    \n",
            "  inflating: data/val/2/28918.png    \n",
            "  inflating: data/val/2/28924.png    \n",
            "  inflating: data/val/2/28942.png    \n",
            "  inflating: data/val/2/28960.png    \n",
            "  inflating: data/val/2/28967.png    \n",
            "  inflating: data/val/2/28973.png    \n",
            "  inflating: data/val/2/28979.png    \n",
            "  inflating: data/val/2/28992.png    \n",
            "  inflating: data/val/2/28998.png    \n",
            "  inflating: data/val/2/29005.png    \n",
            "  inflating: data/val/2/29011.png    \n",
            "  inflating: data/val/2/29021.png    \n",
            "  inflating: data/val/2/29026.png    \n",
            "  inflating: data/val/2/29027.png    \n",
            "  inflating: data/val/2/29030.png    \n",
            "  inflating: data/val/2/29043.png    \n",
            "  inflating: data/val/2/29051.png    \n",
            "  inflating: data/val/2/29057.png    \n",
            "  inflating: data/val/2/29068.png    \n",
            "  inflating: data/val/2/29069.png    \n",
            "  inflating: data/val/2/29070.png    \n",
            "  inflating: data/val/2/29077.png    \n",
            "  inflating: data/val/2/29084.png    \n",
            "  inflating: data/val/2/29092.png    \n",
            "  inflating: data/val/2/29105.png    \n",
            "  inflating: data/val/2/29106.png    \n",
            "  inflating: data/val/2/29107.png    \n",
            "  inflating: data/val/2/29113.png    \n",
            "  inflating: data/val/2/29126.png    \n",
            "  inflating: data/val/2/29146.png    \n",
            "  inflating: data/val/2/29152.png    \n",
            "  inflating: data/val/2/29154.png    \n",
            "  inflating: data/val/2/29157.png    \n",
            "  inflating: data/val/2/29165.png    \n",
            "  inflating: data/val/2/29175.png    \n",
            "  inflating: data/val/2/29178.png    \n",
            "  inflating: data/val/2/29187.png    \n",
            "  inflating: data/val/2/29199.png    \n",
            "  inflating: data/val/2/29200.png    \n",
            "  inflating: data/val/2/29208.png    \n",
            "  inflating: data/val/2/29210.png    \n",
            "  inflating: data/val/2/29219.png    \n",
            "  inflating: data/val/2/29223.png    \n",
            "  inflating: data/val/2/29238.png    \n",
            "  inflating: data/val/2/29249.png    \n",
            "  inflating: data/val/2/29255.png    \n",
            "  inflating: data/val/2/29260.png    \n",
            "  inflating: data/val/2/29262.png    \n",
            "  inflating: data/val/2/29265.png    \n",
            "  inflating: data/val/2/29273.png    \n",
            "  inflating: data/val/2/29277.png    \n",
            "  inflating: data/val/2/29284.png    \n",
            "  inflating: data/val/2/29285.png    \n",
            "  inflating: data/val/2/29287.png    \n",
            "  inflating: data/val/2/29289.png    \n",
            "  inflating: data/val/2/29291.png    \n",
            "  inflating: data/val/2/29294.png    \n",
            "  inflating: data/val/2/29315.png    \n",
            "  inflating: data/val/2/29327.png    \n",
            "  inflating: data/val/2/29328.png    \n",
            "  inflating: data/val/2/29330.png    \n",
            "  inflating: data/val/2/29336.png    \n",
            "  inflating: data/val/2/29346.png    \n",
            "  inflating: data/val/2/29349.png    \n",
            "  inflating: data/val/2/29359.png    \n",
            "  inflating: data/val/2/29360.png    \n",
            "  inflating: data/val/2/29361.png    \n",
            "  inflating: data/val/2/29363.png    \n",
            "  inflating: data/val/2/29364.png    \n",
            "  inflating: data/val/2/29366.png    \n",
            "  inflating: data/val/2/29376.png    \n",
            "  inflating: data/val/2/29382.png    \n",
            "  inflating: data/val/2/29389.png    \n",
            "  inflating: data/val/2/29390.png    \n",
            "  inflating: data/val/2/29405.png    \n",
            "  inflating: data/val/2/29407.png    \n",
            "  inflating: data/val/2/29409.png    \n",
            "  inflating: data/val/2/29411.png    \n",
            "  inflating: data/val/2/29421.png    \n",
            "  inflating: data/val/2/29426.png    \n",
            "  inflating: data/val/2/29429.png    \n",
            "  inflating: data/val/2/29436.png    \n",
            "  inflating: data/val/2/29440.png    \n",
            "  inflating: data/val/2/29444.png    \n",
            "  inflating: data/val/2/29454.png    \n",
            "  inflating: data/val/2/29462.png    \n",
            "  inflating: data/val/2/29473.png    \n",
            "  inflating: data/val/2/29478.png    \n",
            "  inflating: data/val/2/29482.png    \n",
            "  inflating: data/val/2/29488.png    \n",
            "  inflating: data/val/2/29491.png    \n",
            "  inflating: data/val/2/29492.png    \n",
            "  inflating: data/val/2/29496.png    \n",
            "  inflating: data/val/2/29497.png    \n",
            "  inflating: data/val/2/29498.png    \n",
            "  inflating: data/val/2/29513.png    \n",
            "  inflating: data/val/2/29528.png    \n",
            "  inflating: data/val/2/29530.png    \n",
            "  inflating: data/val/2/29532.png    \n",
            "  inflating: data/val/2/29536.png    \n",
            "  inflating: data/val/2/29541.png    \n",
            "  inflating: data/val/2/29543.png    \n",
            "  inflating: data/val/2/29563.png    \n",
            "  inflating: data/val/2/29564.png    \n",
            "  inflating: data/val/2/29567.png    \n",
            "  inflating: data/val/2/29595.png    \n",
            "  inflating: data/val/2/29603.png    \n",
            "  inflating: data/val/2/29610.png    \n",
            "  inflating: data/val/2/29611.png    \n",
            "  inflating: data/val/2/29629.png    \n",
            "  inflating: data/val/2/29655.png    \n",
            "  inflating: data/val/2/29659.png    \n",
            "  inflating: data/val/2/29660.png    \n",
            "  inflating: data/val/2/29662.png    \n",
            "  inflating: data/val/2/29663.png    \n",
            "  inflating: data/val/2/29670.png    \n",
            "  inflating: data/val/2/29674.png    \n",
            "  inflating: data/val/2/29676.png    \n",
            "  inflating: data/val/2/29684.png    \n",
            "  inflating: data/val/2/29688.png    \n",
            "  inflating: data/val/2/29698.png    \n",
            "  inflating: data/val/2/29701.png    \n",
            "  inflating: data/val/2/29705.png    \n",
            "  inflating: data/val/2/29709.png    \n",
            "  inflating: data/val/2/29712.png    \n",
            "  inflating: data/val/2/29713.png    \n",
            "  inflating: data/val/2/29716.png    \n",
            "  inflating: data/val/2/29718.png    \n",
            "  inflating: data/val/2/29720.png    \n",
            "  inflating: data/val/2/29723.png    \n",
            "  inflating: data/val/2/29726.png    \n",
            "  inflating: data/val/2/29731.png    \n",
            "  inflating: data/val/2/29733.png    \n",
            "  inflating: data/val/2/29735.png    \n",
            "  inflating: data/val/2/29736.png    \n",
            "  inflating: data/val/2/29738.png    \n",
            "  inflating: data/val/2/29753.png    \n",
            "  inflating: data/val/2/29762.png    \n",
            "  inflating: data/val/2/29777.png    \n",
            "  inflating: data/val/2/29784.png    \n",
            "  inflating: data/val/2/29793.png    \n",
            "  inflating: data/val/2/29795.png    \n",
            "  inflating: data/val/2/29812.png    \n",
            "  inflating: data/val/2/29813.png    \n",
            "  inflating: data/val/2/29827.png    \n",
            "  inflating: data/val/2/29839.png    \n",
            "  inflating: data/val/2/29846.png    \n",
            "  inflating: data/val/2/29869.png    \n",
            "  inflating: data/val/2/29870.png    \n",
            "  inflating: data/val/2/29871.png    \n",
            "  inflating: data/val/2/29879.png    \n",
            "  inflating: data/val/2/29880.png    \n",
            "  inflating: data/val/2/29883.png    \n",
            "  inflating: data/val/2/29887.png    \n",
            "  inflating: data/val/2/29888.png    \n",
            "  inflating: data/val/2/29892.png    \n",
            "  inflating: data/val/2/29899.png    \n",
            "  inflating: data/val/2/29906.png    \n",
            "  inflating: data/val/2/29908.png    \n",
            "  inflating: data/val/2/29911.png    \n",
            "  inflating: data/val/2/29912.png    \n",
            "  inflating: data/val/2/29918.png    \n",
            "  inflating: data/val/2/29923.png    \n",
            "  inflating: data/val/2/29924.png    \n",
            "  inflating: data/val/2/29925.png    \n",
            "  inflating: data/val/2/29935.png    \n",
            "  inflating: data/val/2/29966.png    \n",
            "  inflating: data/val/2/29983.png    \n",
            "  inflating: data/val/2/29990.png    \n",
            "  inflating: data/val/2/29994.png    \n",
            "  inflating: data/val/2/30013.png    \n",
            "  inflating: data/val/2/30014.png    \n",
            "  inflating: data/val/2/30022.png    \n",
            "  inflating: data/val/2/30027.png    \n",
            "  inflating: data/val/2/30029.png    \n",
            "  inflating: data/val/2/30040.png    \n",
            "  inflating: data/val/2/30043.png    \n",
            "  inflating: data/val/2/30062.png    \n",
            "  inflating: data/val/2/30070.png    \n",
            "  inflating: data/val/2/30074.png    \n",
            "  inflating: data/val/2/30076.png    \n",
            "  inflating: data/val/2/30077.png    \n",
            "  inflating: data/val/2/30078.png    \n",
            "  inflating: data/val/2/30084.png    \n",
            "  inflating: data/val/2/30086.png    \n",
            "  inflating: data/val/2/30115.png    \n",
            "  inflating: data/val/2/30124.png    \n",
            "  inflating: data/val/2/30126.png    \n",
            "  inflating: data/val/2/30129.png    \n",
            "  inflating: data/val/2/30140.png    \n",
            "  inflating: data/val/2/30145.png    \n",
            "  inflating: data/val/2/30151.png    \n",
            "  inflating: data/val/2/30156.png    \n",
            "  inflating: data/val/2/30191.png    \n",
            "  inflating: data/val/2/30196.png    \n",
            "  inflating: data/val/2/30205.png    \n",
            "  inflating: data/val/2/30206.png    \n",
            "  inflating: data/val/2/30207.png    \n",
            "  inflating: data/val/2/30222.png    \n",
            "  inflating: data/val/2/30232.png    \n",
            "  inflating: data/val/2/30237.png    \n",
            "  inflating: data/val/2/30239.png    \n",
            "  inflating: data/val/2/30240.png    \n",
            "  inflating: data/val/2/30242.png    \n",
            "  inflating: data/val/2/30244.png    \n",
            "  inflating: data/val/2/30247.png    \n",
            "  inflating: data/val/2/30250.png    \n",
            "  inflating: data/val/2/30253.png    \n",
            "  inflating: data/val/2/30279.png    \n",
            "  inflating: data/val/2/30296.png    \n",
            "  inflating: data/val/2/30306.png    \n",
            "  inflating: data/val/2/30315.png    \n",
            "  inflating: data/val/2/30316.png    \n",
            "  inflating: data/val/2/30318.png    \n",
            "  inflating: data/val/2/30321.png    \n",
            "  inflating: data/val/2/30324.png    \n",
            "  inflating: data/val/2/30333.png    \n",
            "  inflating: data/val/2/30335.png    \n",
            "  inflating: data/val/2/30349.png    \n",
            "  inflating: data/val/2/30356.png    \n",
            "  inflating: data/val/2/30357.png    \n",
            "  inflating: data/val/2/30360.png    \n",
            "  inflating: data/val/2/30375.png    \n",
            "  inflating: data/val/2/30381.png    \n",
            "  inflating: data/val/2/30396.png    \n",
            "  inflating: data/val/2/30459.png    \n",
            "  inflating: data/val/2/30464.png    \n",
            "  inflating: data/val/2/30465.png    \n",
            "  inflating: data/val/2/30467.png    \n",
            "  inflating: data/val/2/30468.png    \n",
            "  inflating: data/val/2/30475.png    \n",
            "  inflating: data/val/2/30480.png    \n",
            "  inflating: data/val/2/30482.png    \n",
            "  inflating: data/val/2/30491.png    \n",
            "  inflating: data/val/2/30492.png    \n",
            "  inflating: data/val/2/30501.png    \n",
            "  inflating: data/val/2/30524.png    \n",
            "  inflating: data/val/2/30532.png    \n",
            "  inflating: data/val/2/30538.png    \n",
            "  inflating: data/val/2/30549.png    \n",
            "  inflating: data/val/2/30555.png    \n",
            "  inflating: data/val/2/30556.png    \n",
            "  inflating: data/val/2/30558.png    \n",
            "  inflating: data/val/2/30562.png    \n",
            "  inflating: data/val/2/30564.png    \n",
            "  inflating: data/val/2/30570.png    \n",
            "  inflating: data/val/2/30578.png    \n",
            "  inflating: data/val/2/30579.png    \n",
            "  inflating: data/val/2/30594.png    \n",
            "  inflating: data/val/2/30606.png    \n",
            "  inflating: data/val/2/30621.png    \n",
            "  inflating: data/val/2/30629.png    \n",
            "  inflating: data/val/2/30638.png    \n",
            "  inflating: data/val/2/30641.png    \n",
            "  inflating: data/val/2/30657.png    \n",
            "  inflating: data/val/2/30660.png    \n",
            "  inflating: data/val/2/30680.png    \n",
            "  inflating: data/val/2/30687.png    \n",
            "  inflating: data/val/2/30695.png    \n",
            "  inflating: data/val/2/30699.png    \n",
            "  inflating: data/val/2/30706.png    \n",
            "  inflating: data/val/2/30716.png    \n",
            "  inflating: data/val/2/30721.png    \n",
            "  inflating: data/val/2/30723.png    \n",
            "  inflating: data/val/2/30727.png    \n",
            "  inflating: data/val/2/30729.png    \n",
            "  inflating: data/val/2/30732.png    \n",
            "  inflating: data/val/2/30733.png    \n",
            "  inflating: data/val/2/30735.png    \n",
            "  inflating: data/val/2/30744.png    \n",
            "  inflating: data/val/2/30747.png    \n",
            "  inflating: data/val/2/30748.png    \n",
            "  inflating: data/val/2/30753.png    \n",
            "  inflating: data/val/2/30754.png    \n",
            "  inflating: data/val/2/30758.png    \n",
            "  inflating: data/val/2/30770.png    \n",
            "  inflating: data/val/2/30774.png    \n",
            "  inflating: data/val/2/30798.png    \n",
            "  inflating: data/val/2/30799.png    \n",
            "  inflating: data/val/2/30803.png    \n",
            "  inflating: data/val/2/30806.png    \n",
            "  inflating: data/val/2/30813.png    \n",
            "  inflating: data/val/2/30829.png    \n",
            "  inflating: data/val/2/30840.png    \n",
            "  inflating: data/val/2/30857.png    \n",
            "  inflating: data/val/2/30878.png    \n",
            "  inflating: data/val/2/30895.png    \n",
            "  inflating: data/val/2/30896.png    \n",
            "  inflating: data/val/2/30914.png    \n",
            "  inflating: data/val/2/30925.png    \n",
            "  inflating: data/val/2/30936.png    \n",
            "  inflating: data/val/2/30940.png    \n",
            "  inflating: data/val/2/30941.png    \n",
            "  inflating: data/val/2/30947.png    \n",
            "  inflating: data/val/2/30956.png    \n",
            "  inflating: data/val/2/30966.png    \n",
            "  inflating: data/val/2/30983.png    \n",
            "  inflating: data/val/2/31004.png    \n",
            "  inflating: data/val/2/31007.png    \n",
            "  inflating: data/val/2/31014.png    \n",
            "  inflating: data/val/2/31015.png    \n",
            "  inflating: data/val/2/31046.png    \n",
            "  inflating: data/val/2/31079.png    \n",
            "  inflating: data/val/2/31082.png    \n",
            "  inflating: data/val/2/31090.png    \n",
            "  inflating: data/val/2/31093.png    \n",
            "  inflating: data/val/2/31105.png    \n",
            "  inflating: data/val/2/31107.png    \n",
            "  inflating: data/val/2/31114.png    \n",
            "  inflating: data/val/2/31116.png    \n",
            "  inflating: data/val/2/31122.png    \n",
            "  inflating: data/val/2/31130.png    \n",
            "  inflating: data/val/2/31155.png    \n",
            "  inflating: data/val/2/31157.png    \n",
            "  inflating: data/val/2/31159.png    \n",
            "  inflating: data/val/2/31184.png    \n",
            "  inflating: data/val/2/31188.png    \n",
            "  inflating: data/val/2/31201.png    \n",
            "  inflating: data/val/2/31204.png    \n",
            "  inflating: data/val/2/31215.png    \n",
            "  inflating: data/val/2/31228.png    \n",
            "  inflating: data/val/2/31254.png    \n",
            "  inflating: data/val/2/31259.png    \n",
            "  inflating: data/val/2/31262.png    \n",
            "  inflating: data/val/2/31270.png    \n",
            "  inflating: data/val/2/31273.png    \n",
            "  inflating: data/val/2/31286.png    \n",
            "  inflating: data/val/2/31299.png    \n",
            "  inflating: data/val/2/31311.png    \n",
            "  inflating: data/val/2/31312.png    \n",
            "  inflating: data/val/2/31330.png    \n",
            "  inflating: data/val/2/31341.png    \n",
            "  inflating: data/val/2/31342.png    \n",
            "  inflating: data/val/2/31356.png    \n",
            "  inflating: data/val/2/31357.png    \n",
            "  inflating: data/val/2/31360.png    \n",
            "  inflating: data/val/2/31362.png    \n",
            "  inflating: data/val/2/31372.png    \n",
            "  inflating: data/val/2/31375.png    \n",
            "  inflating: data/val/2/31379.png    \n",
            "  inflating: data/val/2/31395.png    \n",
            "  inflating: data/val/2/31396.png    \n",
            "  inflating: data/val/2/31398.png    \n",
            "  inflating: data/val/2/31403.png    \n",
            "  inflating: data/val/2/31411.png    \n",
            "  inflating: data/val/2/31418.png    \n",
            "  inflating: data/val/2/31421.png    \n",
            "  inflating: data/val/2/31426.png    \n",
            "  inflating: data/val/2/31427.png    \n",
            "  inflating: data/val/2/31428.png    \n",
            "  inflating: data/val/2/31441.png    \n",
            "  inflating: data/val/2/31449.png    \n",
            "  inflating: data/val/2/31461.png    \n",
            "  inflating: data/val/2/31465.png    \n",
            "  inflating: data/val/2/31475.png    \n",
            "  inflating: data/val/2/31477.png    \n",
            "  inflating: data/val/2/31481.png    \n",
            "  inflating: data/val/2/31488.png    \n",
            "  inflating: data/val/2/31489.png    \n",
            "  inflating: data/val/2/31492.png    \n",
            "  inflating: data/val/2/31495.png    \n",
            "  inflating: data/val/2/31498.png    \n",
            "  inflating: data/val/2/31513.png    \n",
            "  inflating: data/val/2/31520.png    \n",
            "  inflating: data/val/2/31526.png    \n",
            "  inflating: data/val/2/31535.png    \n",
            "  inflating: data/val/2/31560.png    \n",
            "  inflating: data/val/2/31562.png    \n",
            "  inflating: data/val/2/31566.png    \n",
            "  inflating: data/val/2/31567.png    \n",
            "  inflating: data/val/2/31568.png    \n",
            "  inflating: data/val/2/31569.png    \n",
            "  inflating: data/val/2/31575.png    \n",
            "  inflating: data/val/2/31579.png    \n",
            "  inflating: data/val/2/31584.png    \n",
            "  inflating: data/val/2/31588.png    \n",
            "  inflating: data/val/2/31625.png    \n",
            "  inflating: data/val/2/31630.png    \n",
            "  inflating: data/val/2/31638.png    \n",
            "  inflating: data/val/2/31639.png    \n",
            "  inflating: data/val/2/31655.png    \n",
            "  inflating: data/val/2/31661.png    \n",
            "  inflating: data/val/2/31672.png    \n",
            "  inflating: data/val/2/31678.png    \n",
            "  inflating: data/val/2/31681.png    \n",
            "  inflating: data/val/2/31684.png    \n",
            "  inflating: data/val/2/31702.png    \n",
            "  inflating: data/val/2/31717.png    \n",
            "  inflating: data/val/2/31719.png    \n",
            "  inflating: data/val/2/31732.png    \n",
            "  inflating: data/val/2/31735.png    \n",
            "  inflating: data/val/2/31736.png    \n",
            "  inflating: data/val/2/31738.png    \n",
            "  inflating: data/val/2/31743.png    \n",
            "  inflating: data/val/2/31768.png    \n",
            "  inflating: data/val/2/31771.png    \n",
            "  inflating: data/val/2/31773.png    \n",
            "  inflating: data/val/2/31777.png    \n",
            "  inflating: data/val/2/31780.png    \n",
            "  inflating: data/val/2/31782.png    \n",
            "  inflating: data/val/2/31786.png    \n",
            "  inflating: data/val/2/31797.png    \n",
            "  inflating: data/val/2/31800.png    \n",
            "  inflating: data/val/2/31804.png    \n",
            "  inflating: data/val/2/31818.png    \n",
            "  inflating: data/val/2/31821.png    \n",
            "  inflating: data/val/2/31839.png    \n",
            "  inflating: data/val/2/31842.png    \n",
            "  inflating: data/val/2/31844.png    \n",
            "  inflating: data/val/2/31855.png    \n",
            "  inflating: data/val/2/31860.png    \n",
            "  inflating: data/val/2/31862.png    \n",
            "  inflating: data/val/2/31864.png    \n",
            "  inflating: data/val/2/31866.png    \n",
            "  inflating: data/val/2/31885.png    \n",
            "  inflating: data/val/2/31903.png    \n",
            "  inflating: data/val/2/31904.png    \n",
            "  inflating: data/val/2/31914.png    \n",
            "  inflating: data/val/2/31916.png    \n",
            "  inflating: data/val/2/31919.png    \n",
            "  inflating: data/val/2/31920.png    \n",
            "  inflating: data/val/2/31940.png    \n",
            "  inflating: data/val/2/31944.png    \n",
            "  inflating: data/val/2/31954.png    \n",
            "  inflating: data/val/2/31959.png    \n",
            "  inflating: data/val/2/31965.png    \n",
            "  inflating: data/val/2/31973.png    \n",
            "  inflating: data/val/2/31979.png    \n",
            "  inflating: data/val/2/31982.png    \n",
            "  inflating: data/val/2/31989.png    \n",
            "  inflating: data/val/2/31991.png    \n",
            "  inflating: data/val/2/32000.png    \n",
            "  inflating: data/val/2/32007.png    \n",
            "  inflating: data/val/2/32014.png    \n",
            "  inflating: data/val/2/32018.png    \n",
            "  inflating: data/val/2/32050.png    \n",
            "  inflating: data/val/2/32052.png    \n",
            "  inflating: data/val/2/32053.png    \n",
            "  inflating: data/val/2/32058.png    \n",
            "  inflating: data/val/2/32059.png    \n",
            "  inflating: data/val/2/32074.png    \n",
            "  inflating: data/val/2/32076.png    \n",
            "  inflating: data/val/2/32088.png    \n",
            "  inflating: data/val/2/32093.png    \n",
            "  inflating: data/val/2/32097.png    \n",
            "  inflating: data/val/2/32102.png    \n",
            "  inflating: data/val/2/32103.png    \n",
            "  inflating: data/val/2/32104.png    \n",
            "  inflating: data/val/2/32109.png    \n",
            "  inflating: data/val/2/32111.png    \n",
            "  inflating: data/val/2/32114.png    \n",
            "  inflating: data/val/2/32121.png    \n",
            "  inflating: data/val/2/32125.png    \n",
            "  inflating: data/val/2/32129.png    \n",
            "  inflating: data/val/2/32130.png    \n",
            "  inflating: data/val/2/32135.png    \n",
            "  inflating: data/val/2/32137.png    \n",
            "  inflating: data/val/2/32143.png    \n",
            "  inflating: data/val/2/32168.png    \n",
            "  inflating: data/val/2/32180.png    \n",
            "  inflating: data/val/2/32201.png    \n",
            "  inflating: data/val/2/32205.png    \n",
            "  inflating: data/val/2/32208.png    \n",
            "  inflating: data/val/2/32218.png    \n",
            "  inflating: data/val/2/32229.png    \n",
            "  inflating: data/val/2/32232.png    \n",
            "  inflating: data/val/2/32250.png    \n",
            "  inflating: data/val/2/32271.png    \n",
            "  inflating: data/val/2/32279.png    \n",
            "  inflating: data/val/2/32290.png    \n",
            "  inflating: data/val/3/28713.png    \n",
            "  inflating: data/val/3/28714.png    \n",
            "  inflating: data/val/3/28719.png    \n",
            "  inflating: data/val/3/28728.png    \n",
            "  inflating: data/val/3/28739.png    \n",
            "  inflating: data/val/3/28741.png    \n",
            "  inflating: data/val/3/28748.png    \n",
            "  inflating: data/val/3/28754.png    \n",
            "  inflating: data/val/3/28756.png    \n",
            "  inflating: data/val/3/28760.png    \n",
            "  inflating: data/val/3/28764.png    \n",
            "  inflating: data/val/3/28768.png    \n",
            "  inflating: data/val/3/28770.png    \n",
            "  inflating: data/val/3/28772.png    \n",
            "  inflating: data/val/3/28774.png    \n",
            "  inflating: data/val/3/28792.png    \n",
            "  inflating: data/val/3/28795.png    \n",
            "  inflating: data/val/3/28796.png    \n",
            "  inflating: data/val/3/28797.png    \n",
            "  inflating: data/val/3/28798.png    \n",
            "  inflating: data/val/3/28805.png    \n",
            "  inflating: data/val/3/28806.png    \n",
            "  inflating: data/val/3/28807.png    \n",
            "  inflating: data/val/3/28809.png    \n",
            "  inflating: data/val/3/28812.png    \n",
            "  inflating: data/val/3/28814.png    \n",
            "  inflating: data/val/3/28818.png    \n",
            "  inflating: data/val/3/28819.png    \n",
            "  inflating: data/val/3/28823.png    \n",
            "  inflating: data/val/3/28825.png    \n",
            "  inflating: data/val/3/28835.png    \n",
            "  inflating: data/val/3/28850.png    \n",
            "  inflating: data/val/3/28855.png    \n",
            "  inflating: data/val/3/28856.png    \n",
            "  inflating: data/val/3/28857.png    \n",
            "  inflating: data/val/3/28862.png    \n",
            "  inflating: data/val/3/28868.png    \n",
            "  inflating: data/val/3/28870.png    \n",
            "  inflating: data/val/3/28871.png    \n",
            "  inflating: data/val/3/28875.png    \n",
            "  inflating: data/val/3/28878.png    \n",
            "  inflating: data/val/3/28890.png    \n",
            "  inflating: data/val/3/28892.png    \n",
            "  inflating: data/val/3/28893.png    \n",
            "  inflating: data/val/3/28896.png    \n",
            "  inflating: data/val/3/28897.png    \n",
            "  inflating: data/val/3/28898.png    \n",
            "  inflating: data/val/3/28903.png    \n",
            "  inflating: data/val/3/28909.png    \n",
            "  inflating: data/val/3/28912.png    \n",
            "  inflating: data/val/3/28928.png    \n",
            "  inflating: data/val/3/28929.png    \n",
            "  inflating: data/val/3/28934.png    \n",
            "  inflating: data/val/3/28938.png    \n",
            "  inflating: data/val/3/28943.png    \n",
            "  inflating: data/val/3/28945.png    \n",
            "  inflating: data/val/3/28949.png    \n",
            "  inflating: data/val/3/28954.png    \n",
            "  inflating: data/val/3/28955.png    \n",
            "  inflating: data/val/3/28956.png    \n",
            "  inflating: data/val/3/28962.png    \n",
            "  inflating: data/val/3/28963.png    \n",
            "  inflating: data/val/3/28976.png    \n",
            "  inflating: data/val/3/28977.png    \n",
            "  inflating: data/val/3/28982.png    \n",
            "  inflating: data/val/3/28984.png    \n",
            "  inflating: data/val/3/28989.png    \n",
            "  inflating: data/val/3/28990.png    \n",
            "  inflating: data/val/3/29000.png    \n",
            "  inflating: data/val/3/29003.png    \n",
            "  inflating: data/val/3/29012.png    \n",
            "  inflating: data/val/3/29014.png    \n",
            "  inflating: data/val/3/29016.png    \n",
            "  inflating: data/val/3/29017.png    \n",
            "  inflating: data/val/3/29018.png    \n",
            "  inflating: data/val/3/29019.png    \n",
            "  inflating: data/val/3/29020.png    \n",
            "  inflating: data/val/3/29028.png    \n",
            "  inflating: data/val/3/29031.png    \n",
            "  inflating: data/val/3/29036.png    \n",
            "  inflating: data/val/3/29037.png    \n",
            "  inflating: data/val/3/29039.png    \n",
            "  inflating: data/val/3/29040.png    \n",
            "  inflating: data/val/3/29042.png    \n",
            "  inflating: data/val/3/29044.png    \n",
            "  inflating: data/val/3/29048.png    \n",
            "  inflating: data/val/3/29056.png    \n",
            "  inflating: data/val/3/29062.png    \n",
            "  inflating: data/val/3/29065.png    \n",
            "  inflating: data/val/3/29067.png    \n",
            "  inflating: data/val/3/29071.png    \n",
            "  inflating: data/val/3/29074.png    \n",
            "  inflating: data/val/3/29075.png    \n",
            "  inflating: data/val/3/29076.png    \n",
            "  inflating: data/val/3/29079.png    \n",
            "  inflating: data/val/3/29091.png    \n",
            "  inflating: data/val/3/29093.png    \n",
            "  inflating: data/val/3/29095.png    \n",
            "  inflating: data/val/3/29101.png    \n",
            "  inflating: data/val/3/29111.png    \n",
            "  inflating: data/val/3/29122.png    \n",
            "  inflating: data/val/3/29123.png    \n",
            "  inflating: data/val/3/29124.png    \n",
            "  inflating: data/val/3/29130.png    \n",
            "  inflating: data/val/3/29136.png    \n",
            "  inflating: data/val/3/29137.png    \n",
            "  inflating: data/val/3/29143.png    \n",
            "  inflating: data/val/3/29147.png    \n",
            "  inflating: data/val/3/29160.png    \n",
            "  inflating: data/val/3/29172.png    \n",
            "  inflating: data/val/3/29177.png    \n",
            "  inflating: data/val/3/29179.png    \n",
            "  inflating: data/val/3/29185.png    \n",
            "  inflating: data/val/3/29191.png    \n",
            "  inflating: data/val/3/29197.png    \n",
            "  inflating: data/val/3/29201.png    \n",
            "  inflating: data/val/3/29211.png    \n",
            "  inflating: data/val/3/29214.png    \n",
            "  inflating: data/val/3/29215.png    \n",
            "  inflating: data/val/3/29220.png    \n",
            "  inflating: data/val/3/29224.png    \n",
            "  inflating: data/val/3/29229.png    \n",
            "  inflating: data/val/3/29233.png    \n",
            "  inflating: data/val/3/29234.png    \n",
            "  inflating: data/val/3/29235.png    \n",
            "  inflating: data/val/3/29240.png    \n",
            "  inflating: data/val/3/29248.png    \n",
            "  inflating: data/val/3/29252.png    \n",
            "  inflating: data/val/3/29257.png    \n",
            "  inflating: data/val/3/29258.png    \n",
            "  inflating: data/val/3/29280.png    \n",
            "  inflating: data/val/3/29286.png    \n",
            "  inflating: data/val/3/29290.png    \n",
            "  inflating: data/val/3/29292.png    \n",
            "  inflating: data/val/3/29300.png    \n",
            "  inflating: data/val/3/29303.png    \n",
            "  inflating: data/val/3/29313.png    \n",
            "  inflating: data/val/3/29321.png    \n",
            "  inflating: data/val/3/29329.png    \n",
            "  inflating: data/val/3/29332.png    \n",
            "  inflating: data/val/3/29333.png    \n",
            "  inflating: data/val/3/29334.png    \n",
            "  inflating: data/val/3/29337.png    \n",
            "  inflating: data/val/3/29338.png    \n",
            "  inflating: data/val/3/29340.png    \n",
            "  inflating: data/val/3/29343.png    \n",
            "  inflating: data/val/3/29351.png    \n",
            "  inflating: data/val/3/29353.png    \n",
            "  inflating: data/val/3/29355.png    \n",
            "  inflating: data/val/3/29358.png    \n",
            "  inflating: data/val/3/29362.png    \n",
            "  inflating: data/val/3/29365.png    \n",
            "  inflating: data/val/3/29370.png    \n",
            "  inflating: data/val/3/29371.png    \n",
            "  inflating: data/val/3/29372.png    \n",
            "  inflating: data/val/3/29373.png    \n",
            "  inflating: data/val/3/29396.png    \n",
            "  inflating: data/val/3/29398.png    \n",
            "  inflating: data/val/3/29399.png    \n",
            "  inflating: data/val/3/29404.png    \n",
            "  inflating: data/val/3/29406.png    \n",
            "  inflating: data/val/3/29410.png    \n",
            "  inflating: data/val/3/29415.png    \n",
            "  inflating: data/val/3/29417.png    \n",
            "  inflating: data/val/3/29418.png    \n",
            "  inflating: data/val/3/29422.png    \n",
            "  inflating: data/val/3/29423.png    \n",
            "  inflating: data/val/3/29424.png    \n",
            "  inflating: data/val/3/29427.png    \n",
            "  inflating: data/val/3/29430.png    \n",
            "  inflating: data/val/3/29431.png    \n",
            "  inflating: data/val/3/29439.png    \n",
            "  inflating: data/val/3/29443.png    \n",
            "  inflating: data/val/3/29447.png    \n",
            "  inflating: data/val/3/29448.png    \n",
            "  inflating: data/val/3/29453.png    \n",
            "  inflating: data/val/3/29456.png    \n",
            "  inflating: data/val/3/29458.png    \n",
            "  inflating: data/val/3/29459.png    \n",
            "  inflating: data/val/3/29460.png    \n",
            "  inflating: data/val/3/29464.png    \n",
            "  inflating: data/val/3/29467.png    \n",
            "  inflating: data/val/3/29469.png    \n",
            "  inflating: data/val/3/29472.png    \n",
            "  inflating: data/val/3/29474.png    \n",
            "  inflating: data/val/3/29475.png    \n",
            "  inflating: data/val/3/29479.png    \n",
            "  inflating: data/val/3/29483.png    \n",
            "  inflating: data/val/3/29485.png    \n",
            "  inflating: data/val/3/29495.png    \n",
            "  inflating: data/val/3/29501.png    \n",
            "  inflating: data/val/3/29504.png    \n",
            "  inflating: data/val/3/29509.png    \n",
            "  inflating: data/val/3/29514.png    \n",
            "  inflating: data/val/3/29519.png    \n",
            "  inflating: data/val/3/29526.png    \n",
            "  inflating: data/val/3/29555.png    \n",
            "  inflating: data/val/3/29566.png    \n",
            "  inflating: data/val/3/29577.png    \n",
            "  inflating: data/val/3/29578.png    \n",
            "  inflating: data/val/3/29579.png    \n",
            "  inflating: data/val/3/29587.png    \n",
            "  inflating: data/val/3/29589.png    \n",
            "  inflating: data/val/3/29606.png    \n",
            "  inflating: data/val/3/29612.png    \n",
            "  inflating: data/val/3/29615.png    \n",
            "  inflating: data/val/3/29631.png    \n",
            "  inflating: data/val/3/29634.png    \n",
            "  inflating: data/val/3/29643.png    \n",
            "  inflating: data/val/3/29645.png    \n",
            "  inflating: data/val/3/29647.png    \n",
            "  inflating: data/val/3/29648.png    \n",
            "  inflating: data/val/3/29649.png    \n",
            "  inflating: data/val/3/29650.png    \n",
            "  inflating: data/val/3/29657.png    \n",
            "  inflating: data/val/3/29658.png    \n",
            "  inflating: data/val/3/29668.png    \n",
            "  inflating: data/val/3/29678.png    \n",
            "  inflating: data/val/3/29679.png    \n",
            "  inflating: data/val/3/29680.png    \n",
            "  inflating: data/val/3/29681.png    \n",
            "  inflating: data/val/3/29695.png    \n",
            "  inflating: data/val/3/29696.png    \n",
            "  inflating: data/val/3/29697.png    \n",
            "  inflating: data/val/3/29699.png    \n",
            "  inflating: data/val/3/29707.png    \n",
            "  inflating: data/val/3/29708.png    \n",
            "  inflating: data/val/3/29717.png    \n",
            "  inflating: data/val/3/29722.png    \n",
            "  inflating: data/val/3/29724.png    \n",
            "  inflating: data/val/3/29727.png    \n",
            "  inflating: data/val/3/29740.png    \n",
            "  inflating: data/val/3/29741.png    \n",
            "  inflating: data/val/3/29742.png    \n",
            "  inflating: data/val/3/29744.png    \n",
            "  inflating: data/val/3/29745.png    \n",
            "  inflating: data/val/3/29748.png    \n",
            "  inflating: data/val/3/29751.png    \n",
            "  inflating: data/val/3/29759.png    \n",
            "  inflating: data/val/3/29764.png    \n",
            "  inflating: data/val/3/29766.png    \n",
            "  inflating: data/val/3/29767.png    \n",
            "  inflating: data/val/3/29769.png    \n",
            "  inflating: data/val/3/29770.png    \n",
            "  inflating: data/val/3/29772.png    \n",
            "  inflating: data/val/3/29773.png    \n",
            "  inflating: data/val/3/29776.png    \n",
            "  inflating: data/val/3/29778.png    \n",
            "  inflating: data/val/3/29782.png    \n",
            "  inflating: data/val/3/29789.png    \n",
            "  inflating: data/val/3/29790.png    \n",
            "  inflating: data/val/3/29800.png    \n",
            "  inflating: data/val/3/29801.png    \n",
            "  inflating: data/val/3/29802.png    \n",
            "  inflating: data/val/3/29806.png    \n",
            "  inflating: data/val/3/29808.png    \n",
            "  inflating: data/val/3/29811.png    \n",
            "  inflating: data/val/3/29821.png    \n",
            "  inflating: data/val/3/29826.png    \n",
            "  inflating: data/val/3/29828.png    \n",
            "  inflating: data/val/3/29831.png    \n",
            "  inflating: data/val/3/29834.png    \n",
            "  inflating: data/val/3/29836.png    \n",
            "  inflating: data/val/3/29837.png    \n",
            "  inflating: data/val/3/29842.png    \n",
            "  inflating: data/val/3/29843.png    \n",
            "  inflating: data/val/3/29844.png    \n",
            "  inflating: data/val/3/29847.png    \n",
            "  inflating: data/val/3/29856.png    \n",
            "  inflating: data/val/3/29865.png    \n",
            "  inflating: data/val/3/29867.png    \n",
            "  inflating: data/val/3/29881.png    \n",
            "  inflating: data/val/3/29889.png    \n",
            "  inflating: data/val/3/29894.png    \n",
            "  inflating: data/val/3/29895.png    \n",
            "  inflating: data/val/3/29898.png    \n",
            "  inflating: data/val/3/29903.png    \n",
            "  inflating: data/val/3/29905.png    \n",
            "  inflating: data/val/3/29907.png    \n",
            "  inflating: data/val/3/29910.png    \n",
            "  inflating: data/val/3/29917.png    \n",
            "  inflating: data/val/3/29919.png    \n",
            "  inflating: data/val/3/29926.png    \n",
            "  inflating: data/val/3/29930.png    \n",
            "  inflating: data/val/3/29932.png    \n",
            "  inflating: data/val/3/29934.png    \n",
            "  inflating: data/val/3/29938.png    \n",
            "  inflating: data/val/3/29940.png    \n",
            "  inflating: data/val/3/29945.png    \n",
            "  inflating: data/val/3/29946.png    \n",
            "  inflating: data/val/3/29958.png    \n",
            "  inflating: data/val/3/29959.png    \n",
            "  inflating: data/val/3/29961.png    \n",
            "  inflating: data/val/3/29965.png    \n",
            "  inflating: data/val/3/29967.png    \n",
            "  inflating: data/val/3/29974.png    \n",
            "  inflating: data/val/3/29975.png    \n",
            "  inflating: data/val/3/29979.png    \n",
            "  inflating: data/val/3/29981.png    \n",
            "  inflating: data/val/3/29987.png    \n",
            "  inflating: data/val/3/29988.png    \n",
            "  inflating: data/val/3/29991.png    \n",
            "  inflating: data/val/3/29992.png    \n",
            "  inflating: data/val/3/29993.png    \n",
            "  inflating: data/val/3/29995.png    \n",
            "  inflating: data/val/3/29999.png    \n",
            "  inflating: data/val/3/30003.png    \n",
            "  inflating: data/val/3/30006.png    \n",
            "  inflating: data/val/3/30011.png    \n",
            "  inflating: data/val/3/30016.png    \n",
            "  inflating: data/val/3/30021.png    \n",
            "  inflating: data/val/3/30031.png    \n",
            "  inflating: data/val/3/30033.png    \n",
            "  inflating: data/val/3/30034.png    \n",
            "  inflating: data/val/3/30049.png    \n",
            "  inflating: data/val/3/30052.png    \n",
            "  inflating: data/val/3/30057.png    \n",
            "  inflating: data/val/3/30061.png    \n",
            "  inflating: data/val/3/30064.png    \n",
            "  inflating: data/val/3/30069.png    \n",
            "  inflating: data/val/3/30083.png    \n",
            "  inflating: data/val/3/30089.png    \n",
            "  inflating: data/val/3/30091.png    \n",
            "  inflating: data/val/3/30095.png    \n",
            "  inflating: data/val/3/30097.png    \n",
            "  inflating: data/val/3/30098.png    \n",
            "  inflating: data/val/3/30099.png    \n",
            "  inflating: data/val/3/30101.png    \n",
            "  inflating: data/val/3/30102.png    \n",
            "  inflating: data/val/3/30105.png    \n",
            "  inflating: data/val/3/30106.png    \n",
            "  inflating: data/val/3/30107.png    \n",
            "  inflating: data/val/3/30111.png    \n",
            "  inflating: data/val/3/30114.png    \n",
            "  inflating: data/val/3/30116.png    \n",
            "  inflating: data/val/3/30119.png    \n",
            "  inflating: data/val/3/30127.png    \n",
            "  inflating: data/val/3/30131.png    \n",
            "  inflating: data/val/3/30135.png    \n",
            "  inflating: data/val/3/30143.png    \n",
            "  inflating: data/val/3/30147.png    \n",
            "  inflating: data/val/3/30155.png    \n",
            "  inflating: data/val/3/30158.png    \n",
            "  inflating: data/val/3/30163.png    \n",
            "  inflating: data/val/3/30164.png    \n",
            "  inflating: data/val/3/30166.png    \n",
            "  inflating: data/val/3/30168.png    \n",
            "  inflating: data/val/3/30172.png    \n",
            "  inflating: data/val/3/30175.png    \n",
            "  inflating: data/val/3/30176.png    \n",
            "  inflating: data/val/3/30178.png    \n",
            "  inflating: data/val/3/30180.png    \n",
            "  inflating: data/val/3/30182.png    \n",
            "  inflating: data/val/3/30185.png    \n",
            "  inflating: data/val/3/30186.png    \n",
            "  inflating: data/val/3/30187.png    \n",
            "  inflating: data/val/3/30188.png    \n",
            "  inflating: data/val/3/30193.png    \n",
            "  inflating: data/val/3/30201.png    \n",
            "  inflating: data/val/3/30216.png    \n",
            "  inflating: data/val/3/30217.png    \n",
            "  inflating: data/val/3/30221.png    \n",
            "  inflating: data/val/3/30228.png    \n",
            "  inflating: data/val/3/30236.png    \n",
            "  inflating: data/val/3/30241.png    \n",
            "  inflating: data/val/3/30245.png    \n",
            "  inflating: data/val/3/30248.png    \n",
            "  inflating: data/val/3/30249.png    \n",
            "  inflating: data/val/3/30256.png    \n",
            "  inflating: data/val/3/30261.png    \n",
            "  inflating: data/val/3/30263.png    \n",
            "  inflating: data/val/3/30264.png    \n",
            "  inflating: data/val/3/30265.png    \n",
            "  inflating: data/val/3/30267.png    \n",
            "  inflating: data/val/3/30276.png    \n",
            "  inflating: data/val/3/30280.png    \n",
            "  inflating: data/val/3/30284.png    \n",
            "  inflating: data/val/3/30286.png    \n",
            "  inflating: data/val/3/30288.png    \n",
            "  inflating: data/val/3/30291.png    \n",
            "  inflating: data/val/3/30297.png    \n",
            "  inflating: data/val/3/30299.png    \n",
            "  inflating: data/val/3/30307.png    \n",
            "  inflating: data/val/3/30328.png    \n",
            "  inflating: data/val/3/30329.png    \n",
            "  inflating: data/val/3/30330.png    \n",
            "  inflating: data/val/3/30332.png    \n",
            "  inflating: data/val/3/30334.png    \n",
            "  inflating: data/val/3/30337.png    \n",
            "  inflating: data/val/3/30338.png    \n",
            "  inflating: data/val/3/30341.png    \n",
            "  inflating: data/val/3/30342.png    \n",
            "  inflating: data/val/3/30343.png    \n",
            "  inflating: data/val/3/30348.png    \n",
            "  inflating: data/val/3/30350.png    \n",
            "  inflating: data/val/3/30351.png    \n",
            "  inflating: data/val/3/30359.png    \n",
            "  inflating: data/val/3/30361.png    \n",
            "  inflating: data/val/3/30372.png    \n",
            "  inflating: data/val/3/30380.png    \n",
            "  inflating: data/val/3/30391.png    \n",
            "  inflating: data/val/3/30392.png    \n",
            "  inflating: data/val/3/30400.png    \n",
            "  inflating: data/val/3/30401.png    \n",
            "  inflating: data/val/3/30403.png    \n",
            "  inflating: data/val/3/30406.png    \n",
            "  inflating: data/val/3/30407.png    \n",
            "  inflating: data/val/3/30409.png    \n",
            "  inflating: data/val/3/30414.png    \n",
            "  inflating: data/val/3/30416.png    \n",
            "  inflating: data/val/3/30417.png    \n",
            "  inflating: data/val/3/30424.png    \n",
            "  inflating: data/val/3/30425.png    \n",
            "  inflating: data/val/3/30431.png    \n",
            "  inflating: data/val/3/30436.png    \n",
            "  inflating: data/val/3/30437.png    \n",
            "  inflating: data/val/3/30443.png    \n",
            "  inflating: data/val/3/30449.png    \n",
            "  inflating: data/val/3/30452.png    \n",
            "  inflating: data/val/3/30463.png    \n",
            "  inflating: data/val/3/30471.png    \n",
            "  inflating: data/val/3/30476.png    \n",
            "  inflating: data/val/3/30477.png    \n",
            "  inflating: data/val/3/30478.png    \n",
            "  inflating: data/val/3/30494.png    \n",
            "  inflating: data/val/3/30496.png    \n",
            "  inflating: data/val/3/30498.png    \n",
            "  inflating: data/val/3/30500.png    \n",
            "  inflating: data/val/3/30516.png    \n",
            "  inflating: data/val/3/30517.png    \n",
            "  inflating: data/val/3/30521.png    \n",
            "  inflating: data/val/3/30531.png    \n",
            "  inflating: data/val/3/30533.png    \n",
            "  inflating: data/val/3/30536.png    \n",
            "  inflating: data/val/3/30542.png    \n",
            "  inflating: data/val/3/30554.png    \n",
            "  inflating: data/val/3/30560.png    \n",
            "  inflating: data/val/3/30569.png    \n",
            "  inflating: data/val/3/30571.png    \n",
            "  inflating: data/val/3/30572.png    \n",
            "  inflating: data/val/3/30577.png    \n",
            "  inflating: data/val/3/30580.png    \n",
            "  inflating: data/val/3/30582.png    \n",
            "  inflating: data/val/3/30585.png    \n",
            "  inflating: data/val/3/30587.png    \n",
            "  inflating: data/val/3/30603.png    \n",
            "  inflating: data/val/3/30604.png    \n",
            "  inflating: data/val/3/30608.png    \n",
            "  inflating: data/val/3/30610.png    \n",
            "  inflating: data/val/3/30626.png    \n",
            "  inflating: data/val/3/30635.png    \n",
            "  inflating: data/val/3/30636.png    \n",
            "  inflating: data/val/3/30637.png    \n",
            "  inflating: data/val/3/30646.png    \n",
            "  inflating: data/val/3/30651.png    \n",
            "  inflating: data/val/3/30652.png    \n",
            "  inflating: data/val/3/30661.png    \n",
            "  inflating: data/val/3/30663.png    \n",
            "  inflating: data/val/3/30664.png    \n",
            "  inflating: data/val/3/30667.png    \n",
            "  inflating: data/val/3/30668.png    \n",
            "  inflating: data/val/3/30670.png    \n",
            "  inflating: data/val/3/30672.png    \n",
            "  inflating: data/val/3/30675.png    \n",
            "  inflating: data/val/3/30676.png    \n",
            "  inflating: data/val/3/30678.png    \n",
            "  inflating: data/val/3/30679.png    \n",
            "  inflating: data/val/3/30682.png    \n",
            "  inflating: data/val/3/30683.png    \n",
            "  inflating: data/val/3/30688.png    \n",
            "  inflating: data/val/3/30689.png    \n",
            "  inflating: data/val/3/30693.png    \n",
            "  inflating: data/val/3/30697.png    \n",
            "  inflating: data/val/3/30698.png    \n",
            "  inflating: data/val/3/30700.png    \n",
            "  inflating: data/val/3/30704.png    \n",
            "  inflating: data/val/3/30710.png    \n",
            "  inflating: data/val/3/30720.png    \n",
            "  inflating: data/val/3/30724.png    \n",
            "  inflating: data/val/3/30728.png    \n",
            "  inflating: data/val/3/30730.png    \n",
            "  inflating: data/val/3/30734.png    \n",
            "  inflating: data/val/3/30737.png    \n",
            "  inflating: data/val/3/30738.png    \n",
            "  inflating: data/val/3/30741.png    \n",
            "  inflating: data/val/3/30749.png    \n",
            "  inflating: data/val/3/30755.png    \n",
            "  inflating: data/val/3/30760.png    \n",
            "  inflating: data/val/3/30768.png    \n",
            "  inflating: data/val/3/30771.png    \n",
            "  inflating: data/val/3/30772.png    \n",
            "  inflating: data/val/3/30777.png    \n",
            "  inflating: data/val/3/30779.png    \n",
            "  inflating: data/val/3/30780.png    \n",
            "  inflating: data/val/3/30793.png    \n",
            "  inflating: data/val/3/30795.png    \n",
            "  inflating: data/val/3/30796.png    \n",
            "  inflating: data/val/3/30801.png    \n",
            "  inflating: data/val/3/30802.png    \n",
            "  inflating: data/val/3/30804.png    \n",
            "  inflating: data/val/3/30805.png    \n",
            "  inflating: data/val/3/30807.png    \n",
            "  inflating: data/val/3/30808.png    \n",
            "  inflating: data/val/3/30812.png    \n",
            "  inflating: data/val/3/30814.png    \n",
            "  inflating: data/val/3/30820.png    \n",
            "  inflating: data/val/3/30822.png    \n",
            "  inflating: data/val/3/30823.png    \n",
            "  inflating: data/val/3/30824.png    \n",
            "  inflating: data/val/3/30825.png    \n",
            "  inflating: data/val/3/30826.png    \n",
            "  inflating: data/val/3/30827.png    \n",
            "  inflating: data/val/3/30830.png    \n",
            "  inflating: data/val/3/30832.png    \n",
            "  inflating: data/val/3/30835.png    \n",
            "  inflating: data/val/3/30836.png    \n",
            "  inflating: data/val/3/30837.png    \n",
            "  inflating: data/val/3/30842.png    \n",
            "  inflating: data/val/3/30848.png    \n",
            "  inflating: data/val/3/30849.png    \n",
            "  inflating: data/val/3/30851.png    \n",
            "  inflating: data/val/3/30854.png    \n",
            "  inflating: data/val/3/30855.png    \n",
            "  inflating: data/val/3/30856.png    \n",
            "  inflating: data/val/3/30859.png    \n",
            "  inflating: data/val/3/30861.png    \n",
            "  inflating: data/val/3/30864.png    \n",
            "  inflating: data/val/3/30865.png    \n",
            "  inflating: data/val/3/30870.png    \n",
            "  inflating: data/val/3/30871.png    \n",
            "  inflating: data/val/3/30874.png    \n",
            "  inflating: data/val/3/30875.png    \n",
            "  inflating: data/val/3/30881.png    \n",
            "  inflating: data/val/3/30885.png    \n",
            "  inflating: data/val/3/30891.png    \n",
            "  inflating: data/val/3/30893.png    \n",
            "  inflating: data/val/3/30894.png    \n",
            "  inflating: data/val/3/30897.png    \n",
            "  inflating: data/val/3/30902.png    \n",
            "  inflating: data/val/3/30903.png    \n",
            "  inflating: data/val/3/30904.png    \n",
            "  inflating: data/val/3/30906.png    \n",
            "  inflating: data/val/3/30908.png    \n",
            "  inflating: data/val/3/30909.png    \n",
            "  inflating: data/val/3/30912.png    \n",
            "  inflating: data/val/3/30915.png    \n",
            "  inflating: data/val/3/30917.png    \n",
            "  inflating: data/val/3/30920.png    \n",
            "  inflating: data/val/3/30921.png    \n",
            "  inflating: data/val/3/30938.png    \n",
            "  inflating: data/val/3/30942.png    \n",
            "  inflating: data/val/3/30948.png    \n",
            "  inflating: data/val/3/30951.png    \n",
            "  inflating: data/val/3/30955.png    \n",
            "  inflating: data/val/3/30960.png    \n",
            "  inflating: data/val/3/30961.png    \n",
            "  inflating: data/val/3/30962.png    \n",
            "  inflating: data/val/3/30965.png    \n",
            "  inflating: data/val/3/30970.png    \n",
            "  inflating: data/val/3/30974.png    \n",
            "  inflating: data/val/3/30975.png    \n",
            "  inflating: data/val/3/30979.png    \n",
            "  inflating: data/val/3/30981.png    \n",
            "  inflating: data/val/3/30984.png    \n",
            "  inflating: data/val/3/30986.png    \n",
            "  inflating: data/val/3/30995.png    \n",
            "  inflating: data/val/3/30999.png    \n",
            "  inflating: data/val/3/31000.png    \n",
            "  inflating: data/val/3/31001.png    \n",
            "  inflating: data/val/3/31003.png    \n",
            "  inflating: data/val/3/31006.png    \n",
            "  inflating: data/val/3/31010.png    \n",
            "  inflating: data/val/3/31011.png    \n",
            "  inflating: data/val/3/31012.png    \n",
            "  inflating: data/val/3/31016.png    \n",
            "  inflating: data/val/3/31017.png    \n",
            "  inflating: data/val/3/31018.png    \n",
            "  inflating: data/val/3/31020.png    \n",
            "  inflating: data/val/3/31022.png    \n",
            "  inflating: data/val/3/31024.png    \n",
            "  inflating: data/val/3/31027.png    \n",
            "  inflating: data/val/3/31029.png    \n",
            "  inflating: data/val/3/31035.png    \n",
            "  inflating: data/val/3/31039.png    \n",
            "  inflating: data/val/3/31043.png    \n",
            "  inflating: data/val/3/31048.png    \n",
            "  inflating: data/val/3/31051.png    \n",
            "  inflating: data/val/3/31052.png    \n",
            "  inflating: data/val/3/31053.png    \n",
            "  inflating: data/val/3/31056.png    \n",
            "  inflating: data/val/3/31061.png    \n",
            "  inflating: data/val/3/31062.png    \n",
            "  inflating: data/val/3/31063.png    \n",
            "  inflating: data/val/3/31066.png    \n",
            "  inflating: data/val/3/31074.png    \n",
            "  inflating: data/val/3/31076.png    \n",
            "  inflating: data/val/3/31080.png    \n",
            "  inflating: data/val/3/31084.png    \n",
            "  inflating: data/val/3/31085.png    \n",
            "  inflating: data/val/3/31086.png    \n",
            "  inflating: data/val/3/31087.png    \n",
            "  inflating: data/val/3/31089.png    \n",
            "  inflating: data/val/3/31101.png    \n",
            "  inflating: data/val/3/31106.png    \n",
            "  inflating: data/val/3/31112.png    \n",
            "  inflating: data/val/3/31118.png    \n",
            "  inflating: data/val/3/31119.png    \n",
            "  inflating: data/val/3/31121.png    \n",
            "  inflating: data/val/3/31124.png    \n",
            "  inflating: data/val/3/31128.png    \n",
            "  inflating: data/val/3/31131.png    \n",
            "  inflating: data/val/3/31133.png    \n",
            "  inflating: data/val/3/31134.png    \n",
            "  inflating: data/val/3/31136.png    \n",
            "  inflating: data/val/3/31137.png    \n",
            "  inflating: data/val/3/31140.png    \n",
            "  inflating: data/val/3/31141.png    \n",
            "  inflating: data/val/3/31144.png    \n",
            "  inflating: data/val/3/31145.png    \n",
            "  inflating: data/val/3/31146.png    \n",
            "  inflating: data/val/3/31148.png    \n",
            "  inflating: data/val/3/31150.png    \n",
            "  inflating: data/val/3/31160.png    \n",
            "  inflating: data/val/3/31165.png    \n",
            "  inflating: data/val/3/31167.png    \n",
            "  inflating: data/val/3/31170.png    \n",
            "  inflating: data/val/3/31171.png    \n",
            "  inflating: data/val/3/31176.png    \n",
            "  inflating: data/val/3/31177.png    \n",
            "  inflating: data/val/3/31178.png    \n",
            "  inflating: data/val/3/31179.png    \n",
            "  inflating: data/val/3/31200.png    \n",
            "  inflating: data/val/3/31202.png    \n",
            "  inflating: data/val/3/31208.png    \n",
            "  inflating: data/val/3/31209.png    \n",
            "  inflating: data/val/3/31214.png    \n",
            "  inflating: data/val/3/31230.png    \n",
            "  inflating: data/val/3/31233.png    \n",
            "  inflating: data/val/3/31236.png    \n",
            "  inflating: data/val/3/31239.png    \n",
            "  inflating: data/val/3/31241.png    \n",
            "  inflating: data/val/3/31242.png    \n",
            "  inflating: data/val/3/31243.png    \n",
            "  inflating: data/val/3/31245.png    \n",
            "  inflating: data/val/3/31250.png    \n",
            "  inflating: data/val/3/31251.png    \n",
            "  inflating: data/val/3/31255.png    \n",
            "  inflating: data/val/3/31256.png    \n",
            "  inflating: data/val/3/31258.png    \n",
            "  inflating: data/val/3/31274.png    \n",
            "  inflating: data/val/3/31276.png    \n",
            "  inflating: data/val/3/31282.png    \n",
            "  inflating: data/val/3/31285.png    \n",
            "  inflating: data/val/3/31289.png    \n",
            "  inflating: data/val/3/31291.png    \n",
            "  inflating: data/val/3/31292.png    \n",
            "  inflating: data/val/3/31295.png    \n",
            "  inflating: data/val/3/31297.png    \n",
            "  inflating: data/val/3/31298.png    \n",
            "  inflating: data/val/3/31301.png    \n",
            "  inflating: data/val/3/31303.png    \n",
            "  inflating: data/val/3/31305.png    \n",
            "  inflating: data/val/3/31308.png    \n",
            "  inflating: data/val/3/31309.png    \n",
            "  inflating: data/val/3/31314.png    \n",
            "  inflating: data/val/3/31316.png    \n",
            "  inflating: data/val/3/31320.png    \n",
            "  inflating: data/val/3/31329.png    \n",
            "  inflating: data/val/3/31334.png    \n",
            "  inflating: data/val/3/31335.png    \n",
            "  inflating: data/val/3/31339.png    \n",
            "  inflating: data/val/3/31346.png    \n",
            "  inflating: data/val/3/31349.png    \n",
            "  inflating: data/val/3/31350.png    \n",
            "  inflating: data/val/3/31358.png    \n",
            "  inflating: data/val/3/31359.png    \n",
            "  inflating: data/val/3/31361.png    \n",
            "  inflating: data/val/3/31363.png    \n",
            "  inflating: data/val/3/31366.png    \n",
            "  inflating: data/val/3/31370.png    \n",
            "  inflating: data/val/3/31374.png    \n",
            "  inflating: data/val/3/31376.png    \n",
            "  inflating: data/val/3/31377.png    \n",
            "  inflating: data/val/3/31384.png    \n",
            "  inflating: data/val/3/31385.png    \n",
            "  inflating: data/val/3/31389.png    \n",
            "  inflating: data/val/3/31399.png    \n",
            "  inflating: data/val/3/31402.png    \n",
            "  inflating: data/val/3/31410.png    \n",
            "  inflating: data/val/3/31413.png    \n",
            "  inflating: data/val/3/31419.png    \n",
            "  inflating: data/val/3/31420.png    \n",
            "  inflating: data/val/3/31422.png    \n",
            "  inflating: data/val/3/31438.png    \n",
            "  inflating: data/val/3/31439.png    \n",
            "  inflating: data/val/3/31440.png    \n",
            "  inflating: data/val/3/31452.png    \n",
            "  inflating: data/val/3/31453.png    \n",
            "  inflating: data/val/3/31456.png    \n",
            "  inflating: data/val/3/31457.png    \n",
            "  inflating: data/val/3/31458.png    \n",
            "  inflating: data/val/3/31462.png    \n",
            "  inflating: data/val/3/31474.png    \n",
            "  inflating: data/val/3/31482.png    \n",
            "  inflating: data/val/3/31509.png    \n",
            "  inflating: data/val/3/31514.png    \n",
            "  inflating: data/val/3/31515.png    \n",
            "  inflating: data/val/3/31524.png    \n",
            "  inflating: data/val/3/31525.png    \n",
            "  inflating: data/val/3/31531.png    \n",
            "  inflating: data/val/3/31532.png    \n",
            "  inflating: data/val/3/31540.png    \n",
            "  inflating: data/val/3/31542.png    \n",
            "  inflating: data/val/3/31551.png    \n",
            "  inflating: data/val/3/31554.png    \n",
            "  inflating: data/val/3/31557.png    \n",
            "  inflating: data/val/3/31558.png    \n",
            "  inflating: data/val/3/31571.png    \n",
            "  inflating: data/val/3/31572.png    \n",
            "  inflating: data/val/3/31573.png    \n",
            "  inflating: data/val/3/31574.png    \n",
            "  inflating: data/val/3/31577.png    \n",
            "  inflating: data/val/3/31597.png    \n",
            "  inflating: data/val/3/31598.png    \n",
            "  inflating: data/val/3/31605.png    \n",
            "  inflating: data/val/3/31606.png    \n",
            "  inflating: data/val/3/31617.png    \n",
            "  inflating: data/val/3/31620.png    \n",
            "  inflating: data/val/3/31621.png    \n",
            "  inflating: data/val/3/31627.png    \n",
            "  inflating: data/val/3/31628.png    \n",
            "  inflating: data/val/3/31629.png    \n",
            "  inflating: data/val/3/31636.png    \n",
            "  inflating: data/val/3/31637.png    \n",
            "  inflating: data/val/3/31641.png    \n",
            "  inflating: data/val/3/31644.png    \n",
            "  inflating: data/val/3/31645.png    \n",
            "  inflating: data/val/3/31651.png    \n",
            "  inflating: data/val/3/31659.png    \n",
            "  inflating: data/val/3/31668.png    \n",
            "  inflating: data/val/3/31674.png    \n",
            "  inflating: data/val/3/31675.png    \n",
            "  inflating: data/val/3/31679.png    \n",
            "  inflating: data/val/3/31682.png    \n",
            "  inflating: data/val/3/31683.png    \n",
            "  inflating: data/val/3/31687.png    \n",
            "  inflating: data/val/3/31689.png    \n",
            "  inflating: data/val/3/31691.png    \n",
            "  inflating: data/val/3/31696.png    \n",
            "  inflating: data/val/3/31697.png    \n",
            "  inflating: data/val/3/31707.png    \n",
            "  inflating: data/val/3/31709.png    \n",
            "  inflating: data/val/3/31718.png    \n",
            "  inflating: data/val/3/31723.png    \n",
            "  inflating: data/val/3/31724.png    \n",
            "  inflating: data/val/3/31725.png    \n",
            "  inflating: data/val/3/31727.png    \n",
            "  inflating: data/val/3/31730.png    \n",
            "  inflating: data/val/3/31734.png    \n",
            "  inflating: data/val/3/31737.png    \n",
            "  inflating: data/val/3/31739.png    \n",
            "  inflating: data/val/3/31742.png    \n",
            "  inflating: data/val/3/31746.png    \n",
            "  inflating: data/val/3/31752.png    \n",
            "  inflating: data/val/3/31754.png    \n",
            "  inflating: data/val/3/31755.png    \n",
            "  inflating: data/val/3/31756.png    \n",
            "  inflating: data/val/3/31757.png    \n",
            "  inflating: data/val/3/31760.png    \n",
            "  inflating: data/val/3/31761.png    \n",
            "  inflating: data/val/3/31765.png    \n",
            "  inflating: data/val/3/31770.png    \n",
            "  inflating: data/val/3/31772.png    \n",
            "  inflating: data/val/3/31778.png    \n",
            "  inflating: data/val/3/31785.png    \n",
            "  inflating: data/val/3/31792.png    \n",
            "  inflating: data/val/3/31801.png    \n",
            "  inflating: data/val/3/31802.png    \n",
            "  inflating: data/val/3/31803.png    \n",
            "  inflating: data/val/3/31806.png    \n",
            "  inflating: data/val/3/31810.png    \n",
            "  inflating: data/val/3/31811.png    \n",
            "  inflating: data/val/3/31813.png    \n",
            "  inflating: data/val/3/31822.png    \n",
            "  inflating: data/val/3/31830.png    \n",
            "  inflating: data/val/3/31833.png    \n",
            "  inflating: data/val/3/31846.png    \n",
            "  inflating: data/val/3/31850.png    \n",
            "  inflating: data/val/3/31853.png    \n",
            "  inflating: data/val/3/31856.png    \n",
            "  inflating: data/val/3/31858.png    \n",
            "  inflating: data/val/3/31861.png    \n",
            "  inflating: data/val/3/31863.png    \n",
            "  inflating: data/val/3/31865.png    \n",
            "  inflating: data/val/3/31870.png    \n",
            "  inflating: data/val/3/31883.png    \n",
            "  inflating: data/val/3/31889.png    \n",
            "  inflating: data/val/3/31890.png    \n",
            "  inflating: data/val/3/31896.png    \n",
            "  inflating: data/val/3/31897.png    \n",
            "  inflating: data/val/3/31908.png    \n",
            "  inflating: data/val/3/31910.png    \n",
            "  inflating: data/val/3/31922.png    \n",
            "  inflating: data/val/3/31927.png    \n",
            "  inflating: data/val/3/31929.png    \n",
            "  inflating: data/val/3/31933.png    \n",
            "  inflating: data/val/3/31934.png    \n",
            "  inflating: data/val/3/31945.png    \n",
            "  inflating: data/val/3/31950.png    \n",
            "  inflating: data/val/3/31951.png    \n",
            "  inflating: data/val/3/31952.png    \n",
            "  inflating: data/val/3/31958.png    \n",
            "  inflating: data/val/3/31962.png    \n",
            "  inflating: data/val/3/31964.png    \n",
            "  inflating: data/val/3/31966.png    \n",
            "  inflating: data/val/3/31972.png    \n",
            "  inflating: data/val/3/31977.png    \n",
            "  inflating: data/val/3/31985.png    \n",
            "  inflating: data/val/3/31987.png    \n",
            "  inflating: data/val/3/31990.png    \n",
            "  inflating: data/val/3/32003.png    \n",
            "  inflating: data/val/3/32011.png    \n",
            "  inflating: data/val/3/32012.png    \n",
            "  inflating: data/val/3/32013.png    \n",
            "  inflating: data/val/3/32021.png    \n",
            "  inflating: data/val/3/32022.png    \n",
            "  inflating: data/val/3/32027.png    \n",
            "  inflating: data/val/3/32036.png    \n",
            "  inflating: data/val/3/32037.png    \n",
            "  inflating: data/val/3/32039.png    \n",
            "  inflating: data/val/3/32041.png    \n",
            "  inflating: data/val/3/32043.png    \n",
            "  inflating: data/val/3/32046.png    \n",
            "  inflating: data/val/3/32047.png    \n",
            "  inflating: data/val/3/32048.png    \n",
            "  inflating: data/val/3/32051.png    \n",
            "  inflating: data/val/3/32055.png    \n",
            "  inflating: data/val/3/32062.png    \n",
            "  inflating: data/val/3/32063.png    \n",
            "  inflating: data/val/3/32066.png    \n",
            "  inflating: data/val/3/32068.png    \n",
            "  inflating: data/val/3/32070.png    \n",
            "  inflating: data/val/3/32075.png    \n",
            "  inflating: data/val/3/32080.png    \n",
            "  inflating: data/val/3/32084.png    \n",
            "  inflating: data/val/3/32085.png    \n",
            "  inflating: data/val/3/32087.png    \n",
            "  inflating: data/val/3/32092.png    \n",
            "  inflating: data/val/3/32095.png    \n",
            "  inflating: data/val/3/32100.png    \n",
            "  inflating: data/val/3/32105.png    \n",
            "  inflating: data/val/3/32107.png    \n",
            "  inflating: data/val/3/32108.png    \n",
            "  inflating: data/val/3/32128.png    \n",
            "  inflating: data/val/3/32134.png    \n",
            "  inflating: data/val/3/32136.png    \n",
            "  inflating: data/val/3/32140.png    \n",
            "  inflating: data/val/3/32142.png    \n",
            "  inflating: data/val/3/32144.png    \n",
            "  inflating: data/val/3/32156.png    \n",
            "  inflating: data/val/3/32160.png    \n",
            "  inflating: data/val/3/32162.png    \n",
            "  inflating: data/val/3/32166.png    \n",
            "  inflating: data/val/3/32167.png    \n",
            "  inflating: data/val/3/32169.png    \n",
            "  inflating: data/val/3/32171.png    \n",
            "  inflating: data/val/3/32174.png    \n",
            "  inflating: data/val/3/32177.png    \n",
            "  inflating: data/val/3/32182.png    \n",
            "  inflating: data/val/3/32185.png    \n",
            "  inflating: data/val/3/32186.png    \n",
            "  inflating: data/val/3/32188.png    \n",
            "  inflating: data/val/3/32194.png    \n",
            "  inflating: data/val/3/32198.png    \n",
            "  inflating: data/val/3/32199.png    \n",
            "  inflating: data/val/3/32224.png    \n",
            "  inflating: data/val/3/32227.png    \n",
            "  inflating: data/val/3/32236.png    \n",
            "  inflating: data/val/3/32244.png    \n",
            "  inflating: data/val/3/32247.png    \n",
            "  inflating: data/val/3/32248.png    \n",
            "  inflating: data/val/3/32249.png    \n",
            "  inflating: data/val/3/32251.png    \n",
            "  inflating: data/val/3/32257.png    \n",
            "  inflating: data/val/3/32258.png    \n",
            "  inflating: data/val/3/32259.png    \n",
            "  inflating: data/val/3/32262.png    \n",
            "  inflating: data/val/3/32263.png    \n",
            "  inflating: data/val/3/32273.png    \n",
            "  inflating: data/val/3/32274.png    \n",
            "  inflating: data/val/3/32282.png    \n",
            "  inflating: data/val/3/32287.png    \n",
            "  inflating: data/val/3/32291.png    \n",
            "  inflating: data/val/3/32292.png    \n",
            "  inflating: data/val/3/32294.png    \n",
            "  inflating: data/val/4/28711.png    \n",
            "  inflating: data/val/4/28721.png    \n",
            "  inflating: data/val/4/28732.png    \n",
            "  inflating: data/val/4/28735.png    \n",
            "  inflating: data/val/4/28736.png    \n",
            "  inflating: data/val/4/28744.png    \n",
            "  inflating: data/val/4/28750.png    \n",
            "  inflating: data/val/4/28751.png    \n",
            "  inflating: data/val/4/28762.png    \n",
            "  inflating: data/val/4/28763.png    \n",
            "  inflating: data/val/4/28765.png    \n",
            "  inflating: data/val/4/28771.png    \n",
            "  inflating: data/val/4/28776.png    \n",
            "  inflating: data/val/4/28787.png    \n",
            "  inflating: data/val/4/28788.png    \n",
            "  inflating: data/val/4/28794.png    \n",
            "  inflating: data/val/4/28802.png    \n",
            "  inflating: data/val/4/28833.png    \n",
            "  inflating: data/val/4/28852.png    \n",
            "  inflating: data/val/4/28853.png    \n",
            "  inflating: data/val/4/28865.png    \n",
            "  inflating: data/val/4/28876.png    \n",
            "  inflating: data/val/4/28880.png    \n",
            "  inflating: data/val/4/28886.png    \n",
            "  inflating: data/val/4/28887.png    \n",
            "  inflating: data/val/4/28889.png    \n",
            "  inflating: data/val/4/28895.png    \n",
            "  inflating: data/val/4/28899.png    \n",
            "  inflating: data/val/4/28916.png    \n",
            "  inflating: data/val/4/28921.png    \n",
            "  inflating: data/val/4/28922.png    \n",
            "  inflating: data/val/4/28923.png    \n",
            "  inflating: data/val/4/28926.png    \n",
            "  inflating: data/val/4/28930.png    \n",
            "  inflating: data/val/4/28937.png    \n",
            "  inflating: data/val/4/28944.png    \n",
            "  inflating: data/val/4/28959.png    \n",
            "  inflating: data/val/4/28961.png    \n",
            "  inflating: data/val/4/28966.png    \n",
            "  inflating: data/val/4/28978.png    \n",
            "  inflating: data/val/4/28985.png    \n",
            "  inflating: data/val/4/28993.png    \n",
            "  inflating: data/val/4/29002.png    \n",
            "  inflating: data/val/4/29010.png    \n",
            "  inflating: data/val/4/29025.png    \n",
            "  inflating: data/val/4/29029.png    \n",
            "  inflating: data/val/4/29034.png    \n",
            "  inflating: data/val/4/29046.png    \n",
            "  inflating: data/val/4/29049.png    \n",
            "  inflating: data/val/4/29052.png    \n",
            "  inflating: data/val/4/29058.png    \n",
            "  inflating: data/val/4/29063.png    \n",
            "  inflating: data/val/4/29064.png    \n",
            "  inflating: data/val/4/29078.png    \n",
            "  inflating: data/val/4/29081.png    \n",
            "  inflating: data/val/4/29082.png    \n",
            "  inflating: data/val/4/29083.png    \n",
            "  inflating: data/val/4/29087.png    \n",
            "  inflating: data/val/4/29088.png    \n",
            "  inflating: data/val/4/29097.png    \n",
            "  inflating: data/val/4/29112.png    \n",
            "  inflating: data/val/4/29114.png    \n",
            "  inflating: data/val/4/29121.png    \n",
            "  inflating: data/val/4/29128.png    \n",
            "  inflating: data/val/4/29135.png    \n",
            "  inflating: data/val/4/29138.png    \n",
            "  inflating: data/val/4/29139.png    \n",
            "  inflating: data/val/4/29140.png    \n",
            "  inflating: data/val/4/29141.png    \n",
            "  inflating: data/val/4/29150.png    \n",
            "  inflating: data/val/4/29162.png    \n",
            "  inflating: data/val/4/29163.png    \n",
            "  inflating: data/val/4/29166.png    \n",
            "  inflating: data/val/4/29174.png    \n",
            "  inflating: data/val/4/29188.png    \n",
            "  inflating: data/val/4/29194.png    \n",
            "  inflating: data/val/4/29203.png    \n",
            "  inflating: data/val/4/29204.png    \n",
            "  inflating: data/val/4/29206.png    \n",
            "  inflating: data/val/4/29207.png    \n",
            "  inflating: data/val/4/29218.png    \n",
            "  inflating: data/val/4/29226.png    \n",
            "  inflating: data/val/4/29228.png    \n",
            "  inflating: data/val/4/29231.png    \n",
            "  inflating: data/val/4/29237.png    \n",
            "  inflating: data/val/4/29243.png    \n",
            "  inflating: data/val/4/29251.png    \n",
            "  inflating: data/val/4/29261.png    \n",
            "  inflating: data/val/4/29264.png    \n",
            "  inflating: data/val/4/29270.png    \n",
            "  inflating: data/val/4/29295.png    \n",
            "  inflating: data/val/4/29298.png    \n",
            "  inflating: data/val/4/29301.png    \n",
            "  inflating: data/val/4/29305.png    \n",
            "  inflating: data/val/4/29306.png    \n",
            "  inflating: data/val/4/29307.png    \n",
            "  inflating: data/val/4/29309.png    \n",
            "  inflating: data/val/4/29314.png    \n",
            "  inflating: data/val/4/29316.png    \n",
            "  inflating: data/val/4/29317.png    \n",
            "  inflating: data/val/4/29318.png    \n",
            "  inflating: data/val/4/29322.png    \n",
            "  inflating: data/val/4/29331.png    \n",
            "  inflating: data/val/4/29335.png    \n",
            "  inflating: data/val/4/29342.png    \n",
            "  inflating: data/val/4/29350.png    \n",
            "  inflating: data/val/4/29367.png    \n",
            "  inflating: data/val/4/29369.png    \n",
            "  inflating: data/val/4/29377.png    \n",
            "  inflating: data/val/4/29379.png    \n",
            "  inflating: data/val/4/29381.png    \n",
            "  inflating: data/val/4/29383.png    \n",
            "  inflating: data/val/4/29386.png    \n",
            "  inflating: data/val/4/29387.png    \n",
            "  inflating: data/val/4/29393.png    \n",
            "  inflating: data/val/4/29401.png    \n",
            "  inflating: data/val/4/29412.png    \n",
            "  inflating: data/val/4/29420.png    \n",
            "  inflating: data/val/4/29425.png    \n",
            "  inflating: data/val/4/29433.png    \n",
            "  inflating: data/val/4/29446.png    \n",
            "  inflating: data/val/4/29468.png    \n",
            "  inflating: data/val/4/29481.png    \n",
            "  inflating: data/val/4/29489.png    \n",
            "  inflating: data/val/4/29502.png    \n",
            "  inflating: data/val/4/29510.png    \n",
            "  inflating: data/val/4/29518.png    \n",
            "  inflating: data/val/4/29521.png    \n",
            "  inflating: data/val/4/29524.png    \n",
            "  inflating: data/val/4/29529.png    \n",
            "  inflating: data/val/4/29531.png    \n",
            "  inflating: data/val/4/29538.png    \n",
            "  inflating: data/val/4/29546.png    \n",
            "  inflating: data/val/4/29550.png    \n",
            "  inflating: data/val/4/29552.png    \n",
            "  inflating: data/val/4/29553.png    \n",
            "  inflating: data/val/4/29560.png    \n",
            "  inflating: data/val/4/29561.png    \n",
            "  inflating: data/val/4/29565.png    \n",
            "  inflating: data/val/4/29569.png    \n",
            "  inflating: data/val/4/29573.png    \n",
            "  inflating: data/val/4/29576.png    \n",
            "  inflating: data/val/4/29580.png    \n",
            "  inflating: data/val/4/29583.png    \n",
            "  inflating: data/val/4/29584.png    \n",
            "  inflating: data/val/4/29585.png    \n",
            "  inflating: data/val/4/29594.png    \n",
            "  inflating: data/val/4/29597.png    \n",
            "  inflating: data/val/4/29602.png    \n",
            "  inflating: data/val/4/29617.png    \n",
            "  inflating: data/val/4/29620.png    \n",
            "  inflating: data/val/4/29622.png    \n",
            "  inflating: data/val/4/29623.png    \n",
            "  inflating: data/val/4/29625.png    \n",
            "  inflating: data/val/4/29627.png    \n",
            "  inflating: data/val/4/29628.png    \n",
            "  inflating: data/val/4/29632.png    \n",
            "  inflating: data/val/4/29633.png    \n",
            "  inflating: data/val/4/29635.png    \n",
            "  inflating: data/val/4/29638.png    \n",
            "  inflating: data/val/4/29642.png    \n",
            "  inflating: data/val/4/29644.png    \n",
            "  inflating: data/val/4/29652.png    \n",
            "  inflating: data/val/4/29654.png    \n",
            "  inflating: data/val/4/29665.png    \n",
            "  inflating: data/val/4/29687.png    \n",
            "  inflating: data/val/4/29689.png    \n",
            "  inflating: data/val/4/29692.png    \n",
            "  inflating: data/val/4/29693.png    \n",
            "  inflating: data/val/4/29703.png    \n",
            "  inflating: data/val/4/29710.png    \n",
            "  inflating: data/val/4/29715.png    \n",
            "  inflating: data/val/4/29728.png    \n",
            "  inflating: data/val/4/29737.png    \n",
            "  inflating: data/val/4/29743.png    \n",
            "  inflating: data/val/4/29749.png    \n",
            "  inflating: data/val/4/29755.png    \n",
            "  inflating: data/val/4/29756.png    \n",
            "  inflating: data/val/4/29757.png    \n",
            "  inflating: data/val/4/29761.png    \n",
            "  inflating: data/val/4/29765.png    \n",
            "  inflating: data/val/4/29779.png    \n",
            "  inflating: data/val/4/29780.png    \n",
            "  inflating: data/val/4/29781.png    \n",
            "  inflating: data/val/4/29783.png    \n",
            "  inflating: data/val/4/29786.png    \n",
            "  inflating: data/val/4/29792.png    \n",
            "  inflating: data/val/4/29804.png    \n",
            "  inflating: data/val/4/29816.png    \n",
            "  inflating: data/val/4/29819.png    \n",
            "  inflating: data/val/4/29825.png    \n",
            "  inflating: data/val/4/29832.png    \n",
            "  inflating: data/val/4/29835.png    \n",
            "  inflating: data/val/4/29852.png    \n",
            "  inflating: data/val/4/29854.png    \n",
            "  inflating: data/val/4/29858.png    \n",
            "  inflating: data/val/4/29861.png    \n",
            "  inflating: data/val/4/29873.png    \n",
            "  inflating: data/val/4/29875.png    \n",
            "  inflating: data/val/4/29876.png    \n",
            "  inflating: data/val/4/29897.png    \n",
            "  inflating: data/val/4/29900.png    \n",
            "  inflating: data/val/4/29914.png    \n",
            "  inflating: data/val/4/29922.png    \n",
            "  inflating: data/val/4/29936.png    \n",
            "  inflating: data/val/4/29937.png    \n",
            "  inflating: data/val/4/29941.png    \n",
            "  inflating: data/val/4/29942.png    \n",
            "  inflating: data/val/4/29949.png    \n",
            "  inflating: data/val/4/29956.png    \n",
            "  inflating: data/val/4/29960.png    \n",
            "  inflating: data/val/4/29964.png    \n",
            "  inflating: data/val/4/29970.png    \n",
            "  inflating: data/val/4/29972.png    \n",
            "  inflating: data/val/4/29973.png    \n",
            "  inflating: data/val/4/29982.png    \n",
            "  inflating: data/val/4/29985.png    \n",
            "  inflating: data/val/4/29986.png    \n",
            "  inflating: data/val/4/29998.png    \n",
            "  inflating: data/val/4/30000.png    \n",
            "  inflating: data/val/4/30001.png    \n",
            "  inflating: data/val/4/30017.png    \n",
            "  inflating: data/val/4/30023.png    \n",
            "  inflating: data/val/4/30025.png    \n",
            "  inflating: data/val/4/30032.png    \n",
            "  inflating: data/val/4/30035.png    \n",
            "  inflating: data/val/4/30042.png    \n",
            "  inflating: data/val/4/30045.png    \n",
            "  inflating: data/val/4/30050.png    \n",
            "  inflating: data/val/4/30056.png    \n",
            "  inflating: data/val/4/30080.png    \n",
            "  inflating: data/val/4/30088.png    \n",
            "  inflating: data/val/4/30090.png    \n",
            "  inflating: data/val/4/30094.png    \n",
            "  inflating: data/val/4/30104.png    \n",
            "  inflating: data/val/4/30118.png    \n",
            "  inflating: data/val/4/30123.png    \n",
            "  inflating: data/val/4/30130.png    \n",
            "  inflating: data/val/4/30133.png    \n",
            "  inflating: data/val/4/30134.png    \n",
            "  inflating: data/val/4/30137.png    \n",
            "  inflating: data/val/4/30138.png    \n",
            "  inflating: data/val/4/30144.png    \n",
            "  inflating: data/val/4/30150.png    \n",
            "  inflating: data/val/4/30154.png    \n",
            "  inflating: data/val/4/30162.png    \n",
            "  inflating: data/val/4/30170.png    \n",
            "  inflating: data/val/4/30179.png    \n",
            "  inflating: data/val/4/30181.png    \n",
            "  inflating: data/val/4/30183.png    \n",
            "  inflating: data/val/4/30208.png    \n",
            "  inflating: data/val/4/30211.png    \n",
            "  inflating: data/val/4/30214.png    \n",
            "  inflating: data/val/4/30215.png    \n",
            "  inflating: data/val/4/30219.png    \n",
            "  inflating: data/val/4/30224.png    \n",
            "  inflating: data/val/4/30226.png    \n",
            "  inflating: data/val/4/30231.png    \n",
            "  inflating: data/val/4/30233.png    \n",
            "  inflating: data/val/4/30251.png    \n",
            "  inflating: data/val/4/30254.png    \n",
            "  inflating: data/val/4/30270.png    \n",
            "  inflating: data/val/4/30272.png    \n",
            "  inflating: data/val/4/30274.png    \n",
            "  inflating: data/val/4/30277.png    \n",
            "  inflating: data/val/4/30289.png    \n",
            "  inflating: data/val/4/30292.png    \n",
            "  inflating: data/val/4/30298.png    \n",
            "  inflating: data/val/4/30308.png    \n",
            "  inflating: data/val/4/30310.png    \n",
            "  inflating: data/val/4/30314.png    \n",
            "  inflating: data/val/4/30323.png    \n",
            "  inflating: data/val/4/30326.png    \n",
            "  inflating: data/val/4/30331.png    \n",
            "  inflating: data/val/4/30344.png    \n",
            "  inflating: data/val/4/30358.png    \n",
            "  inflating: data/val/4/30363.png    \n",
            "  inflating: data/val/4/30364.png    \n",
            "  inflating: data/val/4/30377.png    \n",
            "  inflating: data/val/4/30378.png    \n",
            "  inflating: data/val/4/30383.png    \n",
            "  inflating: data/val/4/30389.png    \n",
            "  inflating: data/val/4/30402.png    \n",
            "  inflating: data/val/4/30404.png    \n",
            "  inflating: data/val/4/30410.png    \n",
            "  inflating: data/val/4/30411.png    \n",
            "  inflating: data/val/4/30418.png    \n",
            "  inflating: data/val/4/30421.png    \n",
            "  inflating: data/val/4/30432.png    \n",
            "  inflating: data/val/4/30433.png    \n",
            "  inflating: data/val/4/30434.png    \n",
            "  inflating: data/val/4/30435.png    \n",
            "  inflating: data/val/4/30438.png    \n",
            "  inflating: data/val/4/30439.png    \n",
            "  inflating: data/val/4/30440.png    \n",
            "  inflating: data/val/4/30442.png    \n",
            "  inflating: data/val/4/30444.png    \n",
            "  inflating: data/val/4/30455.png    \n",
            "  inflating: data/val/4/30462.png    \n",
            "  inflating: data/val/4/30466.png    \n",
            "  inflating: data/val/4/30469.png    \n",
            "  inflating: data/val/4/30473.png    \n",
            "  inflating: data/val/4/30487.png    \n",
            "  inflating: data/val/4/30504.png    \n",
            "  inflating: data/val/4/30505.png    \n",
            "  inflating: data/val/4/30511.png    \n",
            "  inflating: data/val/4/30519.png    \n",
            "  inflating: data/val/4/30523.png    \n",
            "  inflating: data/val/4/30525.png    \n",
            "  inflating: data/val/4/30529.png    \n",
            "  inflating: data/val/4/30540.png    \n",
            "  inflating: data/val/4/30541.png    \n",
            "  inflating: data/val/4/30552.png    \n",
            "  inflating: data/val/4/30557.png    \n",
            "  inflating: data/val/4/30565.png    \n",
            "  inflating: data/val/4/30575.png    \n",
            "  inflating: data/val/4/30576.png    \n",
            "  inflating: data/val/4/30588.png    \n",
            "  inflating: data/val/4/30595.png    \n",
            "  inflating: data/val/4/30602.png    \n",
            "  inflating: data/val/4/30607.png    \n",
            "  inflating: data/val/4/30611.png    \n",
            "  inflating: data/val/4/30617.png    \n",
            "  inflating: data/val/4/30632.png    \n",
            "  inflating: data/val/4/30633.png    \n",
            "  inflating: data/val/4/30634.png    \n",
            "  inflating: data/val/4/30645.png    \n",
            "  inflating: data/val/4/30650.png    \n",
            "  inflating: data/val/4/30653.png    \n",
            "  inflating: data/val/4/30654.png    \n",
            "  inflating: data/val/4/30655.png    \n",
            "  inflating: data/val/4/30656.png    \n",
            "  inflating: data/val/4/30659.png    \n",
            "  inflating: data/val/4/30665.png    \n",
            "  inflating: data/val/4/30671.png    \n",
            "  inflating: data/val/4/30674.png    \n",
            "  inflating: data/val/4/30691.png    \n",
            "  inflating: data/val/4/30701.png    \n",
            "  inflating: data/val/4/30702.png    \n",
            "  inflating: data/val/4/30705.png    \n",
            "  inflating: data/val/4/30711.png    \n",
            "  inflating: data/val/4/30712.png    \n",
            "  inflating: data/val/4/30715.png    \n",
            "  inflating: data/val/4/30731.png    \n",
            "  inflating: data/val/4/30740.png    \n",
            "  inflating: data/val/4/30743.png    \n",
            "  inflating: data/val/4/30750.png    \n",
            "  inflating: data/val/4/30752.png    \n",
            "  inflating: data/val/4/30757.png    \n",
            "  inflating: data/val/4/30763.png    \n",
            "  inflating: data/val/4/30764.png    \n",
            "  inflating: data/val/4/30767.png    \n",
            "  inflating: data/val/4/30776.png    \n",
            "  inflating: data/val/4/30784.png    \n",
            "  inflating: data/val/4/30785.png    \n",
            "  inflating: data/val/4/30789.png    \n",
            "  inflating: data/val/4/30791.png    \n",
            "  inflating: data/val/4/30800.png    \n",
            "  inflating: data/val/4/30810.png    \n",
            "  inflating: data/val/4/30816.png    \n",
            "  inflating: data/val/4/30817.png    \n",
            "  inflating: data/val/4/30818.png    \n",
            "  inflating: data/val/4/30833.png    \n",
            "  inflating: data/val/4/30838.png    \n",
            "  inflating: data/val/4/30839.png    \n",
            "  inflating: data/val/4/30846.png    \n",
            "  inflating: data/val/4/30853.png    \n",
            "  inflating: data/val/4/30863.png    \n",
            "  inflating: data/val/4/30867.png    \n",
            "  inflating: data/val/4/30879.png    \n",
            "  inflating: data/val/4/30880.png    \n",
            "  inflating: data/val/4/30892.png    \n",
            "  inflating: data/val/4/30898.png    \n",
            "  inflating: data/val/4/30900.png    \n",
            "  inflating: data/val/4/30911.png    \n",
            "  inflating: data/val/4/30913.png    \n",
            "  inflating: data/val/4/30919.png    \n",
            "  inflating: data/val/4/30922.png    \n",
            "  inflating: data/val/4/30923.png    \n",
            "  inflating: data/val/4/30927.png    \n",
            "  inflating: data/val/4/30946.png    \n",
            "  inflating: data/val/4/30950.png    \n",
            "  inflating: data/val/4/30957.png    \n",
            "  inflating: data/val/4/30958.png    \n",
            "  inflating: data/val/4/30963.png    \n",
            "  inflating: data/val/4/30968.png    \n",
            "  inflating: data/val/4/30969.png    \n",
            "  inflating: data/val/4/30980.png    \n",
            "  inflating: data/val/4/30988.png    \n",
            "  inflating: data/val/4/30991.png    \n",
            "  inflating: data/val/4/30994.png    \n",
            "  inflating: data/val/4/30996.png    \n",
            "  inflating: data/val/4/31002.png    \n",
            "  inflating: data/val/4/31009.png    \n",
            "  inflating: data/val/4/31025.png    \n",
            "  inflating: data/val/4/31032.png    \n",
            "  inflating: data/val/4/31036.png    \n",
            "  inflating: data/val/4/31040.png    \n",
            "  inflating: data/val/4/31049.png    \n",
            "  inflating: data/val/4/31060.png    \n",
            "  inflating: data/val/4/31065.png    \n",
            "  inflating: data/val/4/31067.png    \n",
            "  inflating: data/val/4/31068.png    \n",
            "  inflating: data/val/4/31072.png    \n",
            "  inflating: data/val/4/31078.png    \n",
            "  inflating: data/val/4/31092.png    \n",
            "  inflating: data/val/4/31094.png    \n",
            "  inflating: data/val/4/31099.png    \n",
            "  inflating: data/val/4/31104.png    \n",
            "  inflating: data/val/4/31109.png    \n",
            "  inflating: data/val/4/31111.png    \n",
            "  inflating: data/val/4/31113.png    \n",
            "  inflating: data/val/4/31115.png    \n",
            "  inflating: data/val/4/31117.png    \n",
            "  inflating: data/val/4/31143.png    \n",
            "  inflating: data/val/4/31149.png    \n",
            "  inflating: data/val/4/31156.png    \n",
            "  inflating: data/val/4/31158.png    \n",
            "  inflating: data/val/4/31161.png    \n",
            "  inflating: data/val/4/31166.png    \n",
            "  inflating: data/val/4/31168.png    \n",
            "  inflating: data/val/4/31174.png    \n",
            "  inflating: data/val/4/31175.png    \n",
            "  inflating: data/val/4/31182.png    \n",
            "  inflating: data/val/4/31183.png    \n",
            "  inflating: data/val/4/31186.png    \n",
            "  inflating: data/val/4/31190.png    \n",
            "  inflating: data/val/4/31191.png    \n",
            "  inflating: data/val/4/31197.png    \n",
            "  inflating: data/val/4/31216.png    \n",
            "  inflating: data/val/4/31217.png    \n",
            "  inflating: data/val/4/31218.png    \n",
            "  inflating: data/val/4/31220.png    \n",
            "  inflating: data/val/4/31225.png    \n",
            "  inflating: data/val/4/31227.png    \n",
            "  inflating: data/val/4/31237.png    \n",
            "  inflating: data/val/4/31248.png    \n",
            "  inflating: data/val/4/31249.png    \n",
            "  inflating: data/val/4/31265.png    \n",
            "  inflating: data/val/4/31267.png    \n",
            "  inflating: data/val/4/31272.png    \n",
            "  inflating: data/val/4/31277.png    \n",
            "  inflating: data/val/4/31281.png    \n",
            "  inflating: data/val/4/31283.png    \n",
            "  inflating: data/val/4/31288.png    \n",
            "  inflating: data/val/4/31290.png    \n",
            "  inflating: data/val/4/31300.png    \n",
            "  inflating: data/val/4/31302.png    \n",
            "  inflating: data/val/4/31313.png    \n",
            "  inflating: data/val/4/31319.png    \n",
            "  inflating: data/val/4/31321.png    \n",
            "  inflating: data/val/4/31325.png    \n",
            "  inflating: data/val/4/31326.png    \n",
            "  inflating: data/val/4/31327.png    \n",
            "  inflating: data/val/4/31328.png    \n",
            "  inflating: data/val/4/31333.png    \n",
            "  inflating: data/val/4/31336.png    \n",
            "  inflating: data/val/4/31351.png    \n",
            "  inflating: data/val/4/31355.png    \n",
            "  inflating: data/val/4/31367.png    \n",
            "  inflating: data/val/4/31381.png    \n",
            "  inflating: data/val/4/31386.png    \n",
            "  inflating: data/val/4/31387.png    \n",
            "  inflating: data/val/4/31390.png    \n",
            "  inflating: data/val/4/31393.png    \n",
            "  inflating: data/val/4/31394.png    \n",
            "  inflating: data/val/4/31397.png    \n",
            "  inflating: data/val/4/31401.png    \n",
            "  inflating: data/val/4/31404.png    \n",
            "  inflating: data/val/4/31406.png    \n",
            "  inflating: data/val/4/31407.png    \n",
            "  inflating: data/val/4/31408.png    \n",
            "  inflating: data/val/4/31409.png    \n",
            "  inflating: data/val/4/31416.png    \n",
            "  inflating: data/val/4/31417.png    \n",
            "  inflating: data/val/4/31424.png    \n",
            "  inflating: data/val/4/31431.png    \n",
            "  inflating: data/val/4/31434.png    \n",
            "  inflating: data/val/4/31437.png    \n",
            "  inflating: data/val/4/31454.png    \n",
            "  inflating: data/val/4/31455.png    \n",
            "  inflating: data/val/4/31459.png    \n",
            "  inflating: data/val/4/31460.png    \n",
            "  inflating: data/val/4/31468.png    \n",
            "  inflating: data/val/4/31470.png    \n",
            "  inflating: data/val/4/31471.png    \n",
            "  inflating: data/val/4/31480.png    \n",
            "  inflating: data/val/4/31484.png    \n",
            "  inflating: data/val/4/31490.png    \n",
            "  inflating: data/val/4/31499.png    \n",
            "  inflating: data/val/4/31503.png    \n",
            "  inflating: data/val/4/31506.png    \n",
            "  inflating: data/val/4/31508.png    \n",
            "  inflating: data/val/4/31510.png    \n",
            "  inflating: data/val/4/31511.png    \n",
            "  inflating: data/val/4/31516.png    \n",
            "  inflating: data/val/4/31521.png    \n",
            "  inflating: data/val/4/31522.png    \n",
            "  inflating: data/val/4/31527.png    \n",
            "  inflating: data/val/4/31534.png    \n",
            "  inflating: data/val/4/31536.png    \n",
            "  inflating: data/val/4/31537.png    \n",
            "  inflating: data/val/4/31539.png    \n",
            "  inflating: data/val/4/31547.png    \n",
            "  inflating: data/val/4/31548.png    \n",
            "  inflating: data/val/4/31552.png    \n",
            "  inflating: data/val/4/31555.png    \n",
            "  inflating: data/val/4/31580.png    \n",
            "  inflating: data/val/4/31583.png    \n",
            "  inflating: data/val/4/31585.png    \n",
            "  inflating: data/val/4/31600.png    \n",
            "  inflating: data/val/4/31601.png    \n",
            "  inflating: data/val/4/31610.png    \n",
            "  inflating: data/val/4/31615.png    \n",
            "  inflating: data/val/4/31616.png    \n",
            "  inflating: data/val/4/31619.png    \n",
            "  inflating: data/val/4/31626.png    \n",
            "  inflating: data/val/4/31634.png    \n",
            "  inflating: data/val/4/31642.png    \n",
            "  inflating: data/val/4/31643.png    \n",
            "  inflating: data/val/4/31646.png    \n",
            "  inflating: data/val/4/31652.png    \n",
            "  inflating: data/val/4/31653.png    \n",
            "  inflating: data/val/4/31654.png    \n",
            "  inflating: data/val/4/31656.png    \n",
            "  inflating: data/val/4/31660.png    \n",
            "  inflating: data/val/4/31663.png    \n",
            "  inflating: data/val/4/31665.png    \n",
            "  inflating: data/val/4/31669.png    \n",
            "  inflating: data/val/4/31676.png    \n",
            "  inflating: data/val/4/31700.png    \n",
            "  inflating: data/val/4/31705.png    \n",
            "  inflating: data/val/4/31708.png    \n",
            "  inflating: data/val/4/31710.png    \n",
            "  inflating: data/val/4/31712.png    \n",
            "  inflating: data/val/4/31720.png    \n",
            "  inflating: data/val/4/31721.png    \n",
            "  inflating: data/val/4/31728.png    \n",
            "  inflating: data/val/4/31733.png    \n",
            "  inflating: data/val/4/31740.png    \n",
            "  inflating: data/val/4/31759.png    \n",
            "  inflating: data/val/4/31762.png    \n",
            "  inflating: data/val/4/31783.png    \n",
            "  inflating: data/val/4/31788.png    \n",
            "  inflating: data/val/4/31791.png    \n",
            "  inflating: data/val/4/31807.png    \n",
            "  inflating: data/val/4/31816.png    \n",
            "  inflating: data/val/4/31817.png    \n",
            "  inflating: data/val/4/31819.png    \n",
            "  inflating: data/val/4/31824.png    \n",
            "  inflating: data/val/4/31827.png    \n",
            "  inflating: data/val/4/31829.png    \n",
            "  inflating: data/val/4/31834.png    \n",
            "  inflating: data/val/4/31836.png    \n",
            "  inflating: data/val/4/31837.png    \n",
            "  inflating: data/val/4/31838.png    \n",
            "  inflating: data/val/4/31843.png    \n",
            "  inflating: data/val/4/31845.png    \n",
            "  inflating: data/val/4/31852.png    \n",
            "  inflating: data/val/4/31857.png    \n",
            "  inflating: data/val/4/31859.png    \n",
            "  inflating: data/val/4/31868.png    \n",
            "  inflating: data/val/4/31871.png    \n",
            "  inflating: data/val/4/31881.png    \n",
            "  inflating: data/val/4/31886.png    \n",
            "  inflating: data/val/4/31887.png    \n",
            "  inflating: data/val/4/31893.png    \n",
            "  inflating: data/val/4/31894.png    \n",
            "  inflating: data/val/4/31900.png    \n",
            "  inflating: data/val/4/31901.png    \n",
            "  inflating: data/val/4/31907.png    \n",
            "  inflating: data/val/4/31911.png    \n",
            "  inflating: data/val/4/31912.png    \n",
            "  inflating: data/val/4/31913.png    \n",
            "  inflating: data/val/4/31938.png    \n",
            "  inflating: data/val/4/31941.png    \n",
            "  inflating: data/val/4/31957.png    \n",
            "  inflating: data/val/4/31967.png    \n",
            "  inflating: data/val/4/31975.png    \n",
            "  inflating: data/val/4/31976.png    \n",
            "  inflating: data/val/4/31980.png    \n",
            "  inflating: data/val/4/31984.png    \n",
            "  inflating: data/val/4/31988.png    \n",
            "  inflating: data/val/4/31992.png    \n",
            "  inflating: data/val/4/31995.png    \n",
            "  inflating: data/val/4/32002.png    \n",
            "  inflating: data/val/4/32005.png    \n",
            "  inflating: data/val/4/32006.png    \n",
            "  inflating: data/val/4/32010.png    \n",
            "  inflating: data/val/4/32015.png    \n",
            "  inflating: data/val/4/32025.png    \n",
            "  inflating: data/val/4/32026.png    \n",
            "  inflating: data/val/4/32029.png    \n",
            "  inflating: data/val/4/32030.png    \n",
            "  inflating: data/val/4/32038.png    \n",
            "  inflating: data/val/4/32042.png    \n",
            "  inflating: data/val/4/32044.png    \n",
            "  inflating: data/val/4/32045.png    \n",
            "  inflating: data/val/4/32054.png    \n",
            "  inflating: data/val/4/32060.png    \n",
            "  inflating: data/val/4/32064.png    \n",
            "  inflating: data/val/4/32065.png    \n",
            "  inflating: data/val/4/32067.png    \n",
            "  inflating: data/val/4/32077.png    \n",
            "  inflating: data/val/4/32083.png    \n",
            "  inflating: data/val/4/32089.png    \n",
            "  inflating: data/val/4/32091.png    \n",
            "  inflating: data/val/4/32110.png    \n",
            "  inflating: data/val/4/32124.png    \n",
            "  inflating: data/val/4/32126.png    \n",
            "  inflating: data/val/4/32131.png    \n",
            "  inflating: data/val/4/32138.png    \n",
            "  inflating: data/val/4/32139.png    \n",
            "  inflating: data/val/4/32145.png    \n",
            "  inflating: data/val/4/32147.png    \n",
            "  inflating: data/val/4/32154.png    \n",
            "  inflating: data/val/4/32157.png    \n",
            "  inflating: data/val/4/32163.png    \n",
            "  inflating: data/val/4/32164.png    \n",
            "  inflating: data/val/4/32175.png    \n",
            "  inflating: data/val/4/32183.png    \n",
            "  inflating: data/val/4/32184.png    \n",
            "  inflating: data/val/4/32189.png    \n",
            "  inflating: data/val/4/32190.png    \n",
            "  inflating: data/val/4/32193.png    \n",
            "  inflating: data/val/4/32195.png    \n",
            "  inflating: data/val/4/32197.png    \n",
            "  inflating: data/val/4/32202.png    \n",
            "  inflating: data/val/4/32209.png    \n",
            "  inflating: data/val/4/32217.png    \n",
            "  inflating: data/val/4/32221.png    \n",
            "  inflating: data/val/4/32226.png    \n",
            "  inflating: data/val/4/32233.png    \n",
            "  inflating: data/val/4/32234.png    \n",
            "  inflating: data/val/4/32241.png    \n",
            "  inflating: data/val/4/32245.png    \n",
            "  inflating: data/val/4/32246.png    \n",
            "  inflating: data/val/4/32254.png    \n",
            "  inflating: data/val/4/32255.png    \n",
            "  inflating: data/val/4/32256.png    \n",
            "  inflating: data/val/4/32264.png    \n",
            "  inflating: data/val/4/32267.png    \n",
            "  inflating: data/val/4/32268.png    \n",
            "  inflating: data/val/4/32275.png    \n",
            "  inflating: data/val/4/32276.png    \n",
            "  inflating: data/val/4/32277.png    \n",
            "  inflating: data/val/4/32280.png    \n",
            "  inflating: data/val/4/32286.png    \n",
            "  inflating: data/val/4/32288.png    \n",
            "  inflating: data/val/4/32293.png    \n",
            "  inflating: data/val/4/32295.png    \n",
            "  inflating: data/val/4/32296.png    \n",
            "  inflating: data/val/4/32297.png    \n",
            "  inflating: data/val/5/28725.png    \n",
            "  inflating: data/val/5/28727.png    \n",
            "  inflating: data/val/5/28730.png    \n",
            "  inflating: data/val/5/28755.png    \n",
            "  inflating: data/val/5/28757.png    \n",
            "  inflating: data/val/5/28761.png    \n",
            "  inflating: data/val/5/28767.png    \n",
            "  inflating: data/val/5/28773.png    \n",
            "  inflating: data/val/5/28775.png    \n",
            "  inflating: data/val/5/28780.png    \n",
            "  inflating: data/val/5/28785.png    \n",
            "  inflating: data/val/5/28793.png    \n",
            "  inflating: data/val/5/28803.png    \n",
            "  inflating: data/val/5/28817.png    \n",
            "  inflating: data/val/5/28829.png    \n",
            "  inflating: data/val/5/28837.png    \n",
            "  inflating: data/val/5/28838.png    \n",
            "  inflating: data/val/5/28843.png    \n",
            "  inflating: data/val/5/28844.png    \n",
            "  inflating: data/val/5/28845.png    \n",
            "  inflating: data/val/5/28861.png    \n",
            "  inflating: data/val/5/28863.png    \n",
            "  inflating: data/val/5/28864.png    \n",
            "  inflating: data/val/5/28866.png    \n",
            "  inflating: data/val/5/28877.png    \n",
            "  inflating: data/val/5/28881.png    \n",
            "  inflating: data/val/5/28888.png    \n",
            "  inflating: data/val/5/28891.png    \n",
            "  inflating: data/val/5/28901.png    \n",
            "  inflating: data/val/5/28902.png    \n",
            "  inflating: data/val/5/28905.png    \n",
            "  inflating: data/val/5/28925.png    \n",
            "  inflating: data/val/5/28931.png    \n",
            "  inflating: data/val/5/28935.png    \n",
            "  inflating: data/val/5/28941.png    \n",
            "  inflating: data/val/5/28965.png    \n",
            "  inflating: data/val/5/28981.png    \n",
            "  inflating: data/val/5/28983.png    \n",
            "  inflating: data/val/5/28986.png    \n",
            "  inflating: data/val/5/28999.png    \n",
            "  inflating: data/val/5/29008.png    \n",
            "  inflating: data/val/5/29015.png    \n",
            "  inflating: data/val/5/29038.png    \n",
            "  inflating: data/val/5/29045.png    \n",
            "  inflating: data/val/5/29050.png    \n",
            "  inflating: data/val/5/29073.png    \n",
            "  inflating: data/val/5/29086.png    \n",
            "  inflating: data/val/5/29090.png    \n",
            "  inflating: data/val/5/29096.png    \n",
            "  inflating: data/val/5/29102.png    \n",
            "  inflating: data/val/5/29104.png    \n",
            "  inflating: data/val/5/29110.png    \n",
            "  inflating: data/val/5/29115.png    \n",
            "  inflating: data/val/5/29116.png    \n",
            "  inflating: data/val/5/29117.png    \n",
            "  inflating: data/val/5/29118.png    \n",
            "  inflating: data/val/5/29120.png    \n",
            "  inflating: data/val/5/29127.png    \n",
            "  inflating: data/val/5/29131.png    \n",
            "  inflating: data/val/5/29156.png    \n",
            "  inflating: data/val/5/29164.png    \n",
            "  inflating: data/val/5/29169.png    \n",
            "  inflating: data/val/5/29170.png    \n",
            "  inflating: data/val/5/29180.png    \n",
            "  inflating: data/val/5/29196.png    \n",
            "  inflating: data/val/5/29205.png    \n",
            "  inflating: data/val/5/29212.png    \n",
            "  inflating: data/val/5/29222.png    \n",
            "  inflating: data/val/5/29225.png    \n",
            "  inflating: data/val/5/29236.png    \n",
            "  inflating: data/val/5/29241.png    \n",
            "  inflating: data/val/5/29242.png    \n",
            "  inflating: data/val/5/29247.png    \n",
            "  inflating: data/val/5/29253.png    \n",
            "  inflating: data/val/5/29271.png    \n",
            "  inflating: data/val/5/29274.png    \n",
            "  inflating: data/val/5/29276.png    \n",
            "  inflating: data/val/5/29283.png    \n",
            "  inflating: data/val/5/29288.png    \n",
            "  inflating: data/val/5/29293.png    \n",
            "  inflating: data/val/5/29296.png    \n",
            "  inflating: data/val/5/29320.png    \n",
            "  inflating: data/val/5/29323.png    \n",
            "  inflating: data/val/5/29345.png    \n",
            "  inflating: data/val/5/29354.png    \n",
            "  inflating: data/val/5/29356.png    \n",
            "  inflating: data/val/5/29357.png    \n",
            "  inflating: data/val/5/29368.png    \n",
            "  inflating: data/val/5/29374.png    \n",
            "  inflating: data/val/5/29391.png    \n",
            "  inflating: data/val/5/29400.png    \n",
            "  inflating: data/val/5/29402.png    \n",
            "  inflating: data/val/5/29413.png    \n",
            "  inflating: data/val/5/29457.png    \n",
            "  inflating: data/val/5/29466.png    \n",
            "  inflating: data/val/5/29506.png    \n",
            "  inflating: data/val/5/29508.png    \n",
            "  inflating: data/val/5/29511.png    \n",
            "  inflating: data/val/5/29515.png    \n",
            "  inflating: data/val/5/29516.png    \n",
            "  inflating: data/val/5/29533.png    \n",
            "  inflating: data/val/5/29535.png    \n",
            "  inflating: data/val/5/29537.png    \n",
            "  inflating: data/val/5/29540.png    \n",
            "  inflating: data/val/5/29542.png    \n",
            "  inflating: data/val/5/29551.png    \n",
            "  inflating: data/val/5/29557.png    \n",
            "  inflating: data/val/5/29558.png    \n",
            "  inflating: data/val/5/29568.png    \n",
            "  inflating: data/val/5/29570.png    \n",
            "  inflating: data/val/5/29586.png    \n",
            "  inflating: data/val/5/29593.png    \n",
            "  inflating: data/val/5/29598.png    \n",
            "  inflating: data/val/5/29600.png    \n",
            "  inflating: data/val/5/29613.png    \n",
            "  inflating: data/val/5/29616.png    \n",
            "  inflating: data/val/5/29618.png    \n",
            "  inflating: data/val/5/29619.png    \n",
            "  inflating: data/val/5/29641.png    \n",
            "  inflating: data/val/5/29646.png    \n",
            "  inflating: data/val/5/29672.png    \n",
            "  inflating: data/val/5/29673.png    \n",
            "  inflating: data/val/5/29677.png    \n",
            "  inflating: data/val/5/29690.png    \n",
            "  inflating: data/val/5/29691.png    \n",
            "  inflating: data/val/5/29739.png    \n",
            "  inflating: data/val/5/29747.png    \n",
            "  inflating: data/val/5/29763.png    \n",
            "  inflating: data/val/5/29771.png    \n",
            "  inflating: data/val/5/29791.png    \n",
            "  inflating: data/val/5/29796.png    \n",
            "  inflating: data/val/5/29797.png    \n",
            "  inflating: data/val/5/29798.png    \n",
            "  inflating: data/val/5/29803.png    \n",
            "  inflating: data/val/5/29823.png    \n",
            "  inflating: data/val/5/29824.png    \n",
            "  inflating: data/val/5/29829.png    \n",
            "  inflating: data/val/5/29833.png    \n",
            "  inflating: data/val/5/29845.png    \n",
            "  inflating: data/val/5/29849.png    \n",
            "  inflating: data/val/5/29863.png    \n",
            "  inflating: data/val/5/29864.png    \n",
            "  inflating: data/val/5/29874.png    \n",
            "  inflating: data/val/5/29882.png    \n",
            "  inflating: data/val/5/29884.png    \n",
            "  inflating: data/val/5/29886.png    \n",
            "  inflating: data/val/5/29896.png    \n",
            "  inflating: data/val/5/29904.png    \n",
            "  inflating: data/val/5/29909.png    \n",
            "  inflating: data/val/5/29920.png    \n",
            "  inflating: data/val/5/29927.png    \n",
            "  inflating: data/val/5/29952.png    \n",
            "  inflating: data/val/5/29954.png    \n",
            "  inflating: data/val/5/29955.png    \n",
            "  inflating: data/val/5/29957.png    \n",
            "  inflating: data/val/5/29984.png    \n",
            "  inflating: data/val/5/29989.png    \n",
            "  inflating: data/val/5/29997.png    \n",
            "  inflating: data/val/5/30005.png    \n",
            "  inflating: data/val/5/30008.png    \n",
            "  inflating: data/val/5/30037.png    \n",
            "  inflating: data/val/5/30046.png    \n",
            "  inflating: data/val/5/30048.png    \n",
            "  inflating: data/val/5/30067.png    \n",
            "  inflating: data/val/5/30068.png    \n",
            "  inflating: data/val/5/30071.png    \n",
            "  inflating: data/val/5/30073.png    \n",
            "  inflating: data/val/5/30081.png    \n",
            "  inflating: data/val/5/30087.png    \n",
            "  inflating: data/val/5/30108.png    \n",
            "  inflating: data/val/5/30112.png    \n",
            "  inflating: data/val/5/30136.png    \n",
            "  inflating: data/val/5/30139.png    \n",
            "  inflating: data/val/5/30142.png    \n",
            "  inflating: data/val/5/30152.png    \n",
            "  inflating: data/val/5/30173.png    \n",
            "  inflating: data/val/5/30174.png    \n",
            "  inflating: data/val/5/30177.png    \n",
            "  inflating: data/val/5/30184.png    \n",
            "  inflating: data/val/5/30190.png    \n",
            "  inflating: data/val/5/30192.png    \n",
            "  inflating: data/val/5/30202.png    \n",
            "  inflating: data/val/5/30204.png    \n",
            "  inflating: data/val/5/30209.png    \n",
            "  inflating: data/val/5/30223.png    \n",
            "  inflating: data/val/5/30225.png    \n",
            "  inflating: data/val/5/30243.png    \n",
            "  inflating: data/val/5/30282.png    \n",
            "  inflating: data/val/5/30293.png    \n",
            "  inflating: data/val/5/30300.png    \n",
            "  inflating: data/val/5/30303.png    \n",
            "  inflating: data/val/5/30312.png    \n",
            "  inflating: data/val/5/30322.png    \n",
            "  inflating: data/val/5/30336.png    \n",
            "  inflating: data/val/5/30355.png    \n",
            "  inflating: data/val/5/30368.png    \n",
            "  inflating: data/val/5/30387.png    \n",
            "  inflating: data/val/5/30388.png    \n",
            "  inflating: data/val/5/30390.png    \n",
            "  inflating: data/val/5/30393.png    \n",
            "  inflating: data/val/5/30412.png    \n",
            "  inflating: data/val/5/30423.png    \n",
            "  inflating: data/val/5/30427.png    \n",
            "  inflating: data/val/5/30448.png    \n",
            "  inflating: data/val/5/30451.png    \n",
            "  inflating: data/val/5/30453.png    \n",
            "  inflating: data/val/5/30461.png    \n",
            "  inflating: data/val/5/30489.png    \n",
            "  inflating: data/val/5/30495.png    \n",
            "  inflating: data/val/5/30502.png    \n",
            "  inflating: data/val/5/30507.png    \n",
            "  inflating: data/val/5/30513.png    \n",
            "  inflating: data/val/5/30518.png    \n",
            "  inflating: data/val/5/30520.png    \n",
            "  inflating: data/val/5/30537.png    \n",
            "  inflating: data/val/5/30567.png    \n",
            "  inflating: data/val/5/30568.png    \n",
            "  inflating: data/val/5/30573.png    \n",
            "  inflating: data/val/5/30574.png    \n",
            "  inflating: data/val/5/30586.png    \n",
            "  inflating: data/val/5/30590.png    \n",
            "  inflating: data/val/5/30591.png    \n",
            "  inflating: data/val/5/30596.png    \n",
            "  inflating: data/val/5/30597.png    \n",
            "  inflating: data/val/5/30598.png    \n",
            "  inflating: data/val/5/30612.png    \n",
            "  inflating: data/val/5/30620.png    \n",
            "  inflating: data/val/5/30623.png    \n",
            "  inflating: data/val/5/30625.png    \n",
            "  inflating: data/val/5/30628.png    \n",
            "  inflating: data/val/5/30630.png    \n",
            "  inflating: data/val/5/30640.png    \n",
            "  inflating: data/val/5/30644.png    \n",
            "  inflating: data/val/5/30648.png    \n",
            "  inflating: data/val/5/30658.png    \n",
            "  inflating: data/val/5/30669.png    \n",
            "  inflating: data/val/5/30681.png    \n",
            "  inflating: data/val/5/30714.png    \n",
            "  inflating: data/val/5/30725.png    \n",
            "  inflating: data/val/5/30726.png    \n",
            "  inflating: data/val/5/30739.png    \n",
            "  inflating: data/val/5/30751.png    \n",
            "  inflating: data/val/5/30769.png    \n",
            "  inflating: data/val/5/30778.png    \n",
            "  inflating: data/val/5/30787.png    \n",
            "  inflating: data/val/5/30790.png    \n",
            "  inflating: data/val/5/30792.png    \n",
            "  inflating: data/val/5/30811.png    \n",
            "  inflating: data/val/5/30821.png    \n",
            "  inflating: data/val/5/30862.png    \n",
            "  inflating: data/val/5/30866.png    \n",
            "  inflating: data/val/5/30868.png    \n",
            "  inflating: data/val/5/30873.png    \n",
            "  inflating: data/val/5/30876.png    \n",
            "  inflating: data/val/5/30889.png    \n",
            "  inflating: data/val/5/30899.png    \n",
            "  inflating: data/val/5/30924.png    \n",
            "  inflating: data/val/5/30929.png    \n",
            "  inflating: data/val/5/30935.png    \n",
            "  inflating: data/val/5/30943.png    \n",
            "  inflating: data/val/5/30949.png    \n",
            "  inflating: data/val/5/30964.png    \n",
            "  inflating: data/val/5/30971.png    \n",
            "  inflating: data/val/5/30978.png    \n",
            "  inflating: data/val/5/30989.png    \n",
            "  inflating: data/val/5/30993.png    \n",
            "  inflating: data/val/5/31005.png    \n",
            "  inflating: data/val/5/31008.png    \n",
            "  inflating: data/val/5/31030.png    \n",
            "  inflating: data/val/5/31034.png    \n",
            "  inflating: data/val/5/31037.png    \n",
            "  inflating: data/val/5/31047.png    \n",
            "  inflating: data/val/5/31054.png    \n",
            "  inflating: data/val/5/31055.png    \n",
            "  inflating: data/val/5/31058.png    \n",
            "  inflating: data/val/5/31071.png    \n",
            "  inflating: data/val/5/31091.png    \n",
            "  inflating: data/val/5/31095.png    \n",
            "  inflating: data/val/5/31097.png    \n",
            "  inflating: data/val/5/31102.png    \n",
            "  inflating: data/val/5/31126.png    \n",
            "  inflating: data/val/5/31147.png    \n",
            "  inflating: data/val/5/31163.png    \n",
            "  inflating: data/val/5/31173.png    \n",
            "  inflating: data/val/5/31181.png    \n",
            "  inflating: data/val/5/31185.png    \n",
            "  inflating: data/val/5/31187.png    \n",
            "  inflating: data/val/5/31192.png    \n",
            "  inflating: data/val/5/31194.png    \n",
            "  inflating: data/val/5/31198.png    \n",
            "  inflating: data/val/5/31199.png    \n",
            "  inflating: data/val/5/31206.png    \n",
            "  inflating: data/val/5/31219.png    \n",
            "  inflating: data/val/5/31223.png    \n",
            "  inflating: data/val/5/31224.png    \n",
            "  inflating: data/val/5/31226.png    \n",
            "  inflating: data/val/5/31294.png    \n",
            "  inflating: data/val/5/31304.png    \n",
            "  inflating: data/val/5/31306.png    \n",
            "  inflating: data/val/5/31324.png    \n",
            "  inflating: data/val/5/31340.png    \n",
            "  inflating: data/val/5/31345.png    \n",
            "  inflating: data/val/5/31348.png    \n",
            "  inflating: data/val/5/31353.png    \n",
            "  inflating: data/val/5/31354.png    \n",
            "  inflating: data/val/5/31364.png    \n",
            "  inflating: data/val/5/31391.png    \n",
            "  inflating: data/val/5/31392.png    \n",
            "  inflating: data/val/5/31405.png    \n",
            "  inflating: data/val/5/31423.png    \n",
            "  inflating: data/val/5/31433.png    \n",
            "  inflating: data/val/5/31448.png    \n",
            "  inflating: data/val/5/31451.png    \n",
            "  inflating: data/val/5/31463.png    \n",
            "  inflating: data/val/5/31464.png    \n",
            "  inflating: data/val/5/31473.png    \n",
            "  inflating: data/val/5/31479.png    \n",
            "  inflating: data/val/5/31493.png    \n",
            "  inflating: data/val/5/31496.png    \n",
            "  inflating: data/val/5/31504.png    \n",
            "  inflating: data/val/5/31512.png    \n",
            "  inflating: data/val/5/31523.png    \n",
            "  inflating: data/val/5/31528.png    \n",
            "  inflating: data/val/5/31533.png    \n",
            "  inflating: data/val/5/31545.png    \n",
            "  inflating: data/val/5/31550.png    \n",
            "  inflating: data/val/5/31556.png    \n",
            "  inflating: data/val/5/31559.png    \n",
            "  inflating: data/val/5/31563.png    \n",
            "  inflating: data/val/5/31570.png    \n",
            "  inflating: data/val/5/31587.png    \n",
            "  inflating: data/val/5/31594.png    \n",
            "  inflating: data/val/5/31607.png    \n",
            "  inflating: data/val/5/31612.png    \n",
            "  inflating: data/val/5/31618.png    \n",
            "  inflating: data/val/5/31631.png    \n",
            "  inflating: data/val/5/31632.png    \n",
            "  inflating: data/val/5/31633.png    \n",
            "  inflating: data/val/5/31640.png    \n",
            "  inflating: data/val/5/31662.png    \n",
            "  inflating: data/val/5/31664.png    \n",
            "  inflating: data/val/5/31667.png    \n",
            "  inflating: data/val/5/31671.png    \n",
            "  inflating: data/val/5/31686.png    \n",
            "  inflating: data/val/5/31698.png    \n",
            "  inflating: data/val/5/31699.png    \n",
            "  inflating: data/val/5/31715.png    \n",
            "  inflating: data/val/5/31731.png    \n",
            "  inflating: data/val/5/31741.png    \n",
            "  inflating: data/val/5/31744.png    \n",
            "  inflating: data/val/5/31747.png    \n",
            "  inflating: data/val/5/31758.png    \n",
            "  inflating: data/val/5/31763.png    \n",
            "  inflating: data/val/5/31775.png    \n",
            "  inflating: data/val/5/31781.png    \n",
            "  inflating: data/val/5/31784.png    \n",
            "  inflating: data/val/5/31787.png    \n",
            "  inflating: data/val/5/31790.png    \n",
            "  inflating: data/val/5/31793.png    \n",
            "  inflating: data/val/5/31795.png    \n",
            "  inflating: data/val/5/31799.png    \n",
            "  inflating: data/val/5/31812.png    \n",
            "  inflating: data/val/5/31814.png    \n",
            "  inflating: data/val/5/31826.png    \n",
            "  inflating: data/val/5/31848.png    \n",
            "  inflating: data/val/5/31872.png    \n",
            "  inflating: data/val/5/31875.png    \n",
            "  inflating: data/val/5/31882.png    \n",
            "  inflating: data/val/5/31892.png    \n",
            "  inflating: data/val/5/31895.png    \n",
            "  inflating: data/val/5/31899.png    \n",
            "  inflating: data/val/5/31902.png    \n",
            "  inflating: data/val/5/31905.png    \n",
            "  inflating: data/val/5/31918.png    \n",
            "  inflating: data/val/5/31923.png    \n",
            "  inflating: data/val/5/31935.png    \n",
            "  inflating: data/val/5/31953.png    \n",
            "  inflating: data/val/5/31961.png    \n",
            "  inflating: data/val/5/31968.png    \n",
            "  inflating: data/val/5/31978.png    \n",
            "  inflating: data/val/5/31983.png    \n",
            "  inflating: data/val/5/31998.png    \n",
            "  inflating: data/val/5/32004.png    \n",
            "  inflating: data/val/5/32031.png    \n",
            "  inflating: data/val/5/32035.png    \n",
            "  inflating: data/val/5/32040.png    \n",
            "  inflating: data/val/5/32057.png    \n",
            "  inflating: data/val/5/32061.png    \n",
            "  inflating: data/val/5/32069.png    \n",
            "  inflating: data/val/5/32079.png    \n",
            "  inflating: data/val/5/32096.png    \n",
            "  inflating: data/val/5/32098.png    \n",
            "  inflating: data/val/5/32099.png    \n",
            "  inflating: data/val/5/32119.png    \n",
            "  inflating: data/val/5/32132.png    \n",
            "  inflating: data/val/5/32146.png    \n",
            "  inflating: data/val/5/32148.png    \n",
            "  inflating: data/val/5/32151.png    \n",
            "  inflating: data/val/5/32152.png    \n",
            "  inflating: data/val/5/32159.png    \n",
            "  inflating: data/val/5/32165.png    \n",
            "  inflating: data/val/5/32170.png    \n",
            "  inflating: data/val/5/32173.png    \n",
            "  inflating: data/val/5/32176.png    \n",
            "  inflating: data/val/5/32179.png    \n",
            "  inflating: data/val/5/32210.png    \n",
            "  inflating: data/val/5/32212.png    \n",
            "  inflating: data/val/5/32214.png    \n",
            "  inflating: data/val/5/32235.png    \n",
            "  inflating: data/val/5/32237.png    \n",
            "  inflating: data/val/5/32243.png    \n",
            "  inflating: data/val/5/32265.png    \n",
            "  inflating: data/val/5/32266.png    \n",
            "  inflating: data/val/5/32283.png    \n",
            "  inflating: data/val/5/32284.png    \n",
            "  inflating: data/val/6/28712.png    \n",
            "  inflating: data/val/6/28723.png    \n",
            "  inflating: data/val/6/28740.png    \n",
            "  inflating: data/val/6/28742.png    \n",
            "  inflating: data/val/6/28745.png    \n",
            "  inflating: data/val/6/28747.png    \n",
            "  inflating: data/val/6/28753.png    \n",
            "  inflating: data/val/6/28758.png    \n",
            "  inflating: data/val/6/28769.png    \n",
            "  inflating: data/val/6/28778.png    \n",
            "  inflating: data/val/6/28781.png    \n",
            "  inflating: data/val/6/28782.png    \n",
            "  inflating: data/val/6/28784.png    \n",
            "  inflating: data/val/6/28786.png    \n",
            "  inflating: data/val/6/28799.png    \n",
            "  inflating: data/val/6/28801.png    \n",
            "  inflating: data/val/6/28810.png    \n",
            "  inflating: data/val/6/28815.png    \n",
            "  inflating: data/val/6/28816.png    \n",
            "  inflating: data/val/6/28826.png    \n",
            "  inflating: data/val/6/28828.png    \n",
            "  inflating: data/val/6/28834.png    \n",
            "  inflating: data/val/6/28836.png    \n",
            "  inflating: data/val/6/28842.png    \n",
            "  inflating: data/val/6/28846.png    \n",
            "  inflating: data/val/6/28848.png    \n",
            "  inflating: data/val/6/28849.png    \n",
            "  inflating: data/val/6/28854.png    \n",
            "  inflating: data/val/6/28873.png    \n",
            "  inflating: data/val/6/28874.png    \n",
            "  inflating: data/val/6/28879.png    \n",
            "  inflating: data/val/6/28882.png    \n",
            "  inflating: data/val/6/28900.png    \n",
            "  inflating: data/val/6/28904.png    \n",
            "  inflating: data/val/6/28906.png    \n",
            "  inflating: data/val/6/28910.png    \n",
            "  inflating: data/val/6/28913.png    \n",
            "  inflating: data/val/6/28919.png    \n",
            "  inflating: data/val/6/28936.png    \n",
            "  inflating: data/val/6/28940.png    \n",
            "  inflating: data/val/6/28948.png    \n",
            "  inflating: data/val/6/28950.png    \n",
            "  inflating: data/val/6/28951.png    \n",
            "  inflating: data/val/6/28958.png    \n",
            "  inflating: data/val/6/28964.png    \n",
            "  inflating: data/val/6/28968.png    \n",
            "  inflating: data/val/6/28972.png    \n",
            "  inflating: data/val/6/28980.png    \n",
            "  inflating: data/val/6/28988.png    \n",
            "  inflating: data/val/6/28991.png    \n",
            "  inflating: data/val/6/28997.png    \n",
            "  inflating: data/val/6/29001.png    \n",
            "  inflating: data/val/6/29004.png    \n",
            "  inflating: data/val/6/29006.png    \n",
            "  inflating: data/val/6/29007.png    \n",
            "  inflating: data/val/6/29009.png    \n",
            "  inflating: data/val/6/29022.png    \n",
            "  inflating: data/val/6/29032.png    \n",
            "  inflating: data/val/6/29054.png    \n",
            "  inflating: data/val/6/29055.png    \n",
            "  inflating: data/val/6/29059.png    \n",
            "  inflating: data/val/6/29061.png    \n",
            "  inflating: data/val/6/29072.png    \n",
            "  inflating: data/val/6/29080.png    \n",
            "  inflating: data/val/6/29094.png    \n",
            "  inflating: data/val/6/29099.png    \n",
            "  inflating: data/val/6/29100.png    \n",
            "  inflating: data/val/6/29125.png    \n",
            "  inflating: data/val/6/29129.png    \n",
            "  inflating: data/val/6/29133.png    \n",
            "  inflating: data/val/6/29142.png    \n",
            "  inflating: data/val/6/29144.png    \n",
            "  inflating: data/val/6/29145.png    \n",
            "  inflating: data/val/6/29151.png    \n",
            "  inflating: data/val/6/29161.png    \n",
            "  inflating: data/val/6/29168.png    \n",
            "  inflating: data/val/6/29173.png    \n",
            "  inflating: data/val/6/29176.png    \n",
            "  inflating: data/val/6/29181.png    \n",
            "  inflating: data/val/6/29182.png    \n",
            "  inflating: data/val/6/29183.png    \n",
            "  inflating: data/val/6/29186.png    \n",
            "  inflating: data/val/6/29190.png    \n",
            "  inflating: data/val/6/29193.png    \n",
            "  inflating: data/val/6/29195.png    \n",
            "  inflating: data/val/6/29198.png    \n",
            "  inflating: data/val/6/29209.png    \n",
            "  inflating: data/val/6/29216.png    \n",
            "  inflating: data/val/6/29230.png    \n",
            "  inflating: data/val/6/29232.png    \n",
            "  inflating: data/val/6/29244.png    \n",
            "  inflating: data/val/6/29245.png    \n",
            "  inflating: data/val/6/29246.png    \n",
            "  inflating: data/val/6/29250.png    \n",
            "  inflating: data/val/6/29256.png    \n",
            "  inflating: data/val/6/29268.png    \n",
            "  inflating: data/val/6/29275.png    \n",
            "  inflating: data/val/6/29281.png    \n",
            "  inflating: data/val/6/29297.png    \n",
            "  inflating: data/val/6/29299.png    \n",
            "  inflating: data/val/6/29302.png    \n",
            "  inflating: data/val/6/29304.png    \n",
            "  inflating: data/val/6/29310.png    \n",
            "  inflating: data/val/6/29311.png    \n",
            "  inflating: data/val/6/29312.png    \n",
            "  inflating: data/val/6/29324.png    \n",
            "  inflating: data/val/6/29326.png    \n",
            "  inflating: data/val/6/29339.png    \n",
            "  inflating: data/val/6/29352.png    \n",
            "  inflating: data/val/6/29375.png    \n",
            "  inflating: data/val/6/29380.png    \n",
            "  inflating: data/val/6/29384.png    \n",
            "  inflating: data/val/6/29388.png    \n",
            "  inflating: data/val/6/29392.png    \n",
            "  inflating: data/val/6/29394.png    \n",
            "  inflating: data/val/6/29395.png    \n",
            "  inflating: data/val/6/29397.png    \n",
            "  inflating: data/val/6/29403.png    \n",
            "  inflating: data/val/6/29428.png    \n",
            "  inflating: data/val/6/29437.png    \n",
            "  inflating: data/val/6/29438.png    \n",
            "  inflating: data/val/6/29441.png    \n",
            "  inflating: data/val/6/29442.png    \n",
            "  inflating: data/val/6/29445.png    \n",
            "  inflating: data/val/6/29449.png    \n",
            "  inflating: data/val/6/29452.png    \n",
            "  inflating: data/val/6/29463.png    \n",
            "  inflating: data/val/6/29476.png    \n",
            "  inflating: data/val/6/29484.png    \n",
            "  inflating: data/val/6/29486.png    \n",
            "  inflating: data/val/6/29487.png    \n",
            "  inflating: data/val/6/29490.png    \n",
            "  inflating: data/val/6/29494.png    \n",
            "  inflating: data/val/6/29500.png    \n",
            "  inflating: data/val/6/29505.png    \n",
            "  inflating: data/val/6/29517.png    \n",
            "  inflating: data/val/6/29522.png    \n",
            "  inflating: data/val/6/29525.png    \n",
            "  inflating: data/val/6/29534.png    \n",
            "  inflating: data/val/6/29545.png    \n",
            "  inflating: data/val/6/29549.png    \n",
            "  inflating: data/val/6/29556.png    \n",
            "  inflating: data/val/6/29559.png    \n",
            "  inflating: data/val/6/29562.png    \n",
            "  inflating: data/val/6/29590.png    \n",
            "  inflating: data/val/6/29591.png    \n",
            "  inflating: data/val/6/29592.png    \n",
            "  inflating: data/val/6/29601.png    \n",
            "  inflating: data/val/6/29607.png    \n",
            "  inflating: data/val/6/29608.png    \n",
            "  inflating: data/val/6/29609.png    \n",
            "  inflating: data/val/6/29621.png    \n",
            "  inflating: data/val/6/29624.png    \n",
            "  inflating: data/val/6/29626.png    \n",
            "  inflating: data/val/6/29636.png    \n",
            "  inflating: data/val/6/29637.png    \n",
            "  inflating: data/val/6/29639.png    \n",
            "  inflating: data/val/6/29651.png    \n",
            "  inflating: data/val/6/29656.png    \n",
            "  inflating: data/val/6/29664.png    \n",
            "  inflating: data/val/6/29666.png    \n",
            "  inflating: data/val/6/29667.png    \n",
            "  inflating: data/val/6/29683.png    \n",
            "  inflating: data/val/6/29686.png    \n",
            "  inflating: data/val/6/29694.png    \n",
            "  inflating: data/val/6/29700.png    \n",
            "  inflating: data/val/6/29702.png    \n",
            "  inflating: data/val/6/29704.png    \n",
            "  inflating: data/val/6/29711.png    \n",
            "  inflating: data/val/6/29714.png    \n",
            "  inflating: data/val/6/29721.png    \n",
            "  inflating: data/val/6/29725.png    \n",
            "  inflating: data/val/6/29729.png    \n",
            "  inflating: data/val/6/29750.png    \n",
            "  inflating: data/val/6/29752.png    \n",
            "  inflating: data/val/6/29754.png    \n",
            "  inflating: data/val/6/29758.png    \n",
            "  inflating: data/val/6/29760.png    \n",
            "  inflating: data/val/6/29774.png    \n",
            "  inflating: data/val/6/29775.png    \n",
            "  inflating: data/val/6/29787.png    \n",
            "  inflating: data/val/6/29794.png    \n",
            "  inflating: data/val/6/29805.png    \n",
            "  inflating: data/val/6/29809.png    \n",
            "  inflating: data/val/6/29820.png    \n",
            "  inflating: data/val/6/29822.png    \n",
            "  inflating: data/val/6/29840.png    \n",
            "  inflating: data/val/6/29841.png    \n",
            "  inflating: data/val/6/29848.png    \n",
            "  inflating: data/val/6/29853.png    \n",
            "  inflating: data/val/6/29857.png    \n",
            "  inflating: data/val/6/29859.png    \n",
            "  inflating: data/val/6/29872.png    \n",
            "  inflating: data/val/6/29877.png    \n",
            "  inflating: data/val/6/29885.png    \n",
            "  inflating: data/val/6/29890.png    \n",
            "  inflating: data/val/6/29891.png    \n",
            "  inflating: data/val/6/29893.png    \n",
            "  inflating: data/val/6/29902.png    \n",
            "  inflating: data/val/6/29913.png    \n",
            "  inflating: data/val/6/29915.png    \n",
            "  inflating: data/val/6/29916.png    \n",
            "  inflating: data/val/6/29921.png    \n",
            "  inflating: data/val/6/29928.png    \n",
            "  inflating: data/val/6/29939.png    \n",
            "  inflating: data/val/6/29943.png    \n",
            "  inflating: data/val/6/29944.png    \n",
            "  inflating: data/val/6/29948.png    \n",
            "  inflating: data/val/6/29953.png    \n",
            "  inflating: data/val/6/29968.png    \n",
            "  inflating: data/val/6/29969.png    \n",
            "  inflating: data/val/6/29978.png    \n",
            "  inflating: data/val/6/29996.png    \n",
            "  inflating: data/val/6/30007.png    \n",
            "  inflating: data/val/6/30010.png    \n",
            "  inflating: data/val/6/30012.png    \n",
            "  inflating: data/val/6/30018.png    \n",
            "  inflating: data/val/6/30019.png    \n",
            "  inflating: data/val/6/30020.png    \n",
            "  inflating: data/val/6/30030.png    \n",
            "  inflating: data/val/6/30036.png    \n",
            "  inflating: data/val/6/30038.png    \n",
            "  inflating: data/val/6/30039.png    \n",
            "  inflating: data/val/6/30041.png    \n",
            "  inflating: data/val/6/30044.png    \n",
            "  inflating: data/val/6/30051.png    \n",
            "  inflating: data/val/6/30053.png    \n",
            "  inflating: data/val/6/30054.png    \n",
            "  inflating: data/val/6/30055.png    \n",
            "  inflating: data/val/6/30060.png    \n",
            "  inflating: data/val/6/30063.png    \n",
            "  inflating: data/val/6/30065.png    \n",
            "  inflating: data/val/6/30072.png    \n",
            "  inflating: data/val/6/30075.png    \n",
            "  inflating: data/val/6/30096.png    \n",
            "  inflating: data/val/6/30100.png    \n",
            "  inflating: data/val/6/30103.png    \n",
            "  inflating: data/val/6/30109.png    \n",
            "  inflating: data/val/6/30110.png    \n",
            "  inflating: data/val/6/30117.png    \n",
            "  inflating: data/val/6/30121.png    \n",
            "  inflating: data/val/6/30122.png    \n",
            "  inflating: data/val/6/30125.png    \n",
            "  inflating: data/val/6/30146.png    \n",
            "  inflating: data/val/6/30148.png    \n",
            "  inflating: data/val/6/30149.png    \n",
            "  inflating: data/val/6/30157.png    \n",
            "  inflating: data/val/6/30159.png    \n",
            "  inflating: data/val/6/30160.png    \n",
            "  inflating: data/val/6/30161.png    \n",
            "  inflating: data/val/6/30167.png    \n",
            "  inflating: data/val/6/30169.png    \n",
            "  inflating: data/val/6/30189.png    \n",
            "  inflating: data/val/6/30195.png    \n",
            "  inflating: data/val/6/30197.png    \n",
            "  inflating: data/val/6/30198.png    \n",
            "  inflating: data/val/6/30199.png    \n",
            "  inflating: data/val/6/30200.png    \n",
            "  inflating: data/val/6/30203.png    \n",
            "  inflating: data/val/6/30210.png    \n",
            "  inflating: data/val/6/30212.png    \n",
            "  inflating: data/val/6/30218.png    \n",
            "  inflating: data/val/6/30220.png    \n",
            "  inflating: data/val/6/30229.png    \n",
            "  inflating: data/val/6/30234.png    \n",
            "  inflating: data/val/6/30238.png    \n",
            "  inflating: data/val/6/30255.png    \n",
            "  inflating: data/val/6/30257.png    \n",
            "  inflating: data/val/6/30258.png    \n",
            "  inflating: data/val/6/30266.png    \n",
            "  inflating: data/val/6/30269.png    \n",
            "  inflating: data/val/6/30271.png    \n",
            "  inflating: data/val/6/30275.png    \n",
            "  inflating: data/val/6/30278.png    \n",
            "  inflating: data/val/6/30281.png    \n",
            "  inflating: data/val/6/30283.png    \n",
            "  inflating: data/val/6/30287.png    \n",
            "  inflating: data/val/6/30294.png    \n",
            "  inflating: data/val/6/30295.png    \n",
            "  inflating: data/val/6/30301.png    \n",
            "  inflating: data/val/6/30304.png    \n",
            "  inflating: data/val/6/30311.png    \n",
            "  inflating: data/val/6/30313.png    \n",
            "  inflating: data/val/6/30319.png    \n",
            "  inflating: data/val/6/30339.png    \n",
            "  inflating: data/val/6/30345.png    \n",
            "  inflating: data/val/6/30346.png    \n",
            "  inflating: data/val/6/30352.png    \n",
            "  inflating: data/val/6/30353.png    \n",
            "  inflating: data/val/6/30354.png    \n",
            "  inflating: data/val/6/30362.png    \n",
            "  inflating: data/val/6/30365.png    \n",
            "  inflating: data/val/6/30366.png    \n",
            "  inflating: data/val/6/30367.png    \n",
            "  inflating: data/val/6/30370.png    \n",
            "  inflating: data/val/6/30371.png    \n",
            "  inflating: data/val/6/30373.png    \n",
            "  inflating: data/val/6/30374.png    \n",
            "  inflating: data/val/6/30376.png    \n",
            "  inflating: data/val/6/30379.png    \n",
            "  inflating: data/val/6/30385.png    \n",
            "  inflating: data/val/6/30386.png    \n",
            "  inflating: data/val/6/30395.png    \n",
            "  inflating: data/val/6/30397.png    \n",
            "  inflating: data/val/6/30408.png    \n",
            "  inflating: data/val/6/30419.png    \n",
            "  inflating: data/val/6/30422.png    \n",
            "  inflating: data/val/6/30426.png    \n",
            "  inflating: data/val/6/30429.png    \n",
            "  inflating: data/val/6/30441.png    \n",
            "  inflating: data/val/6/30450.png    \n",
            "  inflating: data/val/6/30454.png    \n",
            "  inflating: data/val/6/30457.png    \n",
            "  inflating: data/val/6/30472.png    \n",
            "  inflating: data/val/6/30474.png    \n",
            "  inflating: data/val/6/30481.png    \n",
            "  inflating: data/val/6/30483.png    \n",
            "  inflating: data/val/6/30485.png    \n",
            "  inflating: data/val/6/30488.png    \n",
            "  inflating: data/val/6/30510.png    \n",
            "  inflating: data/val/6/30512.png    \n",
            "  inflating: data/val/6/30515.png    \n",
            "  inflating: data/val/6/30522.png    \n",
            "  inflating: data/val/6/30527.png    \n",
            "  inflating: data/val/6/30528.png    \n",
            "  inflating: data/val/6/30530.png    \n",
            "  inflating: data/val/6/30534.png    \n",
            "  inflating: data/val/6/30539.png    \n",
            "  inflating: data/val/6/30543.png    \n",
            "  inflating: data/val/6/30546.png    \n",
            "  inflating: data/val/6/30547.png    \n",
            "  inflating: data/val/6/30551.png    \n",
            "  inflating: data/val/6/30553.png    \n",
            "  inflating: data/val/6/30559.png    \n",
            "  inflating: data/val/6/30561.png    \n",
            "  inflating: data/val/6/30566.png    \n",
            "  inflating: data/val/6/30581.png    \n",
            "  inflating: data/val/6/30583.png    \n",
            "  inflating: data/val/6/30600.png    \n",
            "  inflating: data/val/6/30613.png    \n",
            "  inflating: data/val/6/30614.png    \n",
            "  inflating: data/val/6/30615.png    \n",
            "  inflating: data/val/6/30616.png    \n",
            "  inflating: data/val/6/30618.png    \n",
            "  inflating: data/val/6/30622.png    \n",
            "  inflating: data/val/6/30624.png    \n",
            "  inflating: data/val/6/30642.png    \n",
            "  inflating: data/val/6/30643.png    \n",
            "  inflating: data/val/6/30647.png    \n",
            "  inflating: data/val/6/30662.png    \n",
            "  inflating: data/val/6/30666.png    \n",
            "  inflating: data/val/6/30677.png    \n",
            "  inflating: data/val/6/30684.png    \n",
            "  inflating: data/val/6/30686.png    \n",
            "  inflating: data/val/6/30692.png    \n",
            "  inflating: data/val/6/30703.png    \n",
            "  inflating: data/val/6/30707.png    \n",
            "  inflating: data/val/6/30708.png    \n",
            "  inflating: data/val/6/30713.png    \n",
            "  inflating: data/val/6/30717.png    \n",
            "  inflating: data/val/6/30718.png    \n",
            "  inflating: data/val/6/30722.png    \n",
            "  inflating: data/val/6/30736.png    \n",
            "  inflating: data/val/6/30762.png    \n",
            "  inflating: data/val/6/30765.png    \n",
            "  inflating: data/val/6/30766.png    \n",
            "  inflating: data/val/6/30775.png    \n",
            "  inflating: data/val/6/30782.png    \n",
            "  inflating: data/val/6/30786.png    \n",
            "  inflating: data/val/6/30788.png    \n",
            "  inflating: data/val/6/30794.png    \n",
            "  inflating: data/val/6/30809.png    \n",
            "  inflating: data/val/6/30819.png    \n",
            "  inflating: data/val/6/30841.png    \n",
            "  inflating: data/val/6/30843.png    \n",
            "  inflating: data/val/6/30844.png    \n",
            "  inflating: data/val/6/30850.png    \n",
            "  inflating: data/val/6/30860.png    \n",
            "  inflating: data/val/6/30869.png    \n",
            "  inflating: data/val/6/30877.png    \n",
            "  inflating: data/val/6/30882.png    \n",
            "  inflating: data/val/6/30883.png    \n",
            "  inflating: data/val/6/30884.png    \n",
            "  inflating: data/val/6/30888.png    \n",
            "  inflating: data/val/6/30910.png    \n",
            "  inflating: data/val/6/30916.png    \n",
            "  inflating: data/val/6/30918.png    \n",
            "  inflating: data/val/6/30926.png    \n",
            "  inflating: data/val/6/30932.png    \n",
            "  inflating: data/val/6/30934.png    \n",
            "  inflating: data/val/6/30937.png    \n",
            "  inflating: data/val/6/30939.png    \n",
            "  inflating: data/val/6/30945.png    \n",
            "  inflating: data/val/6/30953.png    \n",
            "  inflating: data/val/6/30954.png    \n",
            "  inflating: data/val/6/30959.png    \n",
            "  inflating: data/val/6/30973.png    \n",
            "  inflating: data/val/6/30976.png    \n",
            "  inflating: data/val/6/30985.png    \n",
            "  inflating: data/val/6/30987.png    \n",
            "  inflating: data/val/6/30990.png    \n",
            "  inflating: data/val/6/30992.png    \n",
            "  inflating: data/val/6/30997.png    \n",
            "  inflating: data/val/6/30998.png    \n",
            "  inflating: data/val/6/31021.png    \n",
            "  inflating: data/val/6/31023.png    \n",
            "  inflating: data/val/6/31026.png    \n",
            "  inflating: data/val/6/31038.png    \n",
            "  inflating: data/val/6/31041.png    \n",
            "  inflating: data/val/6/31050.png    \n",
            "  inflating: data/val/6/31057.png    \n",
            "  inflating: data/val/6/31059.png    \n",
            "  inflating: data/val/6/31070.png    \n",
            "  inflating: data/val/6/31073.png    \n",
            "  inflating: data/val/6/31075.png    \n",
            "  inflating: data/val/6/31096.png    \n",
            "  inflating: data/val/6/31110.png    \n",
            "  inflating: data/val/6/31123.png    \n",
            "  inflating: data/val/6/31127.png    \n",
            "  inflating: data/val/6/31129.png    \n",
            "  inflating: data/val/6/31132.png    \n",
            "  inflating: data/val/6/31135.png    \n",
            "  inflating: data/val/6/31139.png    \n",
            "  inflating: data/val/6/31164.png    \n",
            "  inflating: data/val/6/31169.png    \n",
            "  inflating: data/val/6/31172.png    \n",
            "  inflating: data/val/6/31180.png    \n",
            "  inflating: data/val/6/31189.png    \n",
            "  inflating: data/val/6/31211.png    \n",
            "  inflating: data/val/6/31212.png    \n",
            "  inflating: data/val/6/31238.png    \n",
            "  inflating: data/val/6/31240.png    \n",
            "  inflating: data/val/6/31244.png    \n",
            "  inflating: data/val/6/31246.png    \n",
            "  inflating: data/val/6/31252.png    \n",
            "  inflating: data/val/6/31257.png    \n",
            "  inflating: data/val/6/31260.png    \n",
            "  inflating: data/val/6/31263.png    \n",
            "  inflating: data/val/6/31264.png    \n",
            "  inflating: data/val/6/31268.png    \n",
            "  inflating: data/val/6/31269.png    \n",
            "  inflating: data/val/6/31275.png    \n",
            "  inflating: data/val/6/31278.png    \n",
            "  inflating: data/val/6/31284.png    \n",
            "  inflating: data/val/6/31287.png    \n",
            "  inflating: data/val/6/31307.png    \n",
            "  inflating: data/val/6/31315.png    \n",
            "  inflating: data/val/6/31317.png    \n",
            "  inflating: data/val/6/31318.png    \n",
            "  inflating: data/val/6/31332.png    \n",
            "  inflating: data/val/6/31344.png    \n",
            "  inflating: data/val/6/31347.png    \n",
            "  inflating: data/val/6/31352.png    \n",
            "  inflating: data/val/6/31365.png    \n",
            "  inflating: data/val/6/31369.png    \n",
            "  inflating: data/val/6/31371.png    \n",
            "  inflating: data/val/6/31373.png    \n",
            "  inflating: data/val/6/31380.png    \n",
            "  inflating: data/val/6/31382.png    \n",
            "  inflating: data/val/6/31412.png    \n",
            "  inflating: data/val/6/31414.png    \n",
            "  inflating: data/val/6/31415.png    \n",
            "  inflating: data/val/6/31425.png    \n",
            "  inflating: data/val/6/31432.png    \n",
            "  inflating: data/val/6/31435.png    \n",
            "  inflating: data/val/6/31436.png    \n",
            "  inflating: data/val/6/31443.png    \n",
            "  inflating: data/val/6/31446.png    \n",
            "  inflating: data/val/6/31466.png    \n",
            "  inflating: data/val/6/31467.png    \n",
            "  inflating: data/val/6/31469.png    \n",
            "  inflating: data/val/6/31478.png    \n",
            "  inflating: data/val/6/31483.png    \n",
            "  inflating: data/val/6/31486.png    \n",
            "  inflating: data/val/6/31487.png    \n",
            "  inflating: data/val/6/31501.png    \n",
            "  inflating: data/val/6/31502.png    \n",
            "  inflating: data/val/6/31507.png    \n",
            "  inflating: data/val/6/31517.png    \n",
            "  inflating: data/val/6/31541.png    \n",
            "  inflating: data/val/6/31546.png    \n",
            "  inflating: data/val/6/31553.png    \n",
            "  inflating: data/val/6/31561.png    \n",
            "  inflating: data/val/6/31565.png    \n",
            "  inflating: data/val/6/31576.png    \n",
            "  inflating: data/val/6/31578.png    \n",
            "  inflating: data/val/6/31582.png    \n",
            "  inflating: data/val/6/31591.png    \n",
            "  inflating: data/val/6/31593.png    \n",
            "  inflating: data/val/6/31596.png    \n",
            "  inflating: data/val/6/31602.png    \n",
            "  inflating: data/val/6/31608.png    \n",
            "  inflating: data/val/6/31614.png    \n",
            "  inflating: data/val/6/31623.png    \n",
            "  inflating: data/val/6/31624.png    \n",
            "  inflating: data/val/6/31647.png    \n",
            "  inflating: data/val/6/31649.png    \n",
            "  inflating: data/val/6/31650.png    \n",
            "  inflating: data/val/6/31658.png    \n",
            "  inflating: data/val/6/31677.png    \n",
            "  inflating: data/val/6/31680.png    \n",
            "  inflating: data/val/6/31685.png    \n",
            "  inflating: data/val/6/31688.png    \n",
            "  inflating: data/val/6/31690.png    \n",
            "  inflating: data/val/6/31695.png    \n",
            "  inflating: data/val/6/31701.png    \n",
            "  inflating: data/val/6/31704.png    \n",
            "  inflating: data/val/6/31713.png    \n",
            "  inflating: data/val/6/31714.png    \n",
            "  inflating: data/val/6/31745.png    \n",
            "  inflating: data/val/6/31748.png    \n",
            "  inflating: data/val/6/31750.png    \n",
            "  inflating: data/val/6/31751.png    \n",
            "  inflating: data/val/6/31753.png    \n",
            "  inflating: data/val/6/31764.png    \n",
            "  inflating: data/val/6/31766.png    \n",
            "  inflating: data/val/6/31767.png    \n",
            "  inflating: data/val/6/31769.png    \n",
            "  inflating: data/val/6/31774.png    \n",
            "  inflating: data/val/6/31776.png    \n",
            "  inflating: data/val/6/31789.png    \n",
            "  inflating: data/val/6/31798.png    \n",
            "  inflating: data/val/6/31805.png    \n",
            "  inflating: data/val/6/31808.png    \n",
            "  inflating: data/val/6/31820.png    \n",
            "  inflating: data/val/6/31823.png    \n",
            "  inflating: data/val/6/31825.png    \n",
            "  inflating: data/val/6/31831.png    \n",
            "  inflating: data/val/6/31832.png    \n",
            "  inflating: data/val/6/31835.png    \n",
            "  inflating: data/val/6/31841.png    \n",
            "  inflating: data/val/6/31847.png    \n",
            "  inflating: data/val/6/31851.png    \n",
            "  inflating: data/val/6/31854.png    \n",
            "  inflating: data/val/6/31869.png    \n",
            "  inflating: data/val/6/31876.png    \n",
            "  inflating: data/val/6/31878.png    \n",
            "  inflating: data/val/6/31880.png    \n",
            "  inflating: data/val/6/31884.png    \n",
            "  inflating: data/val/6/31891.png    \n",
            "  inflating: data/val/6/31898.png    \n",
            "  inflating: data/val/6/31906.png    \n",
            "  inflating: data/val/6/31909.png    \n",
            "  inflating: data/val/6/31915.png    \n",
            "  inflating: data/val/6/31924.png    \n",
            "  inflating: data/val/6/31930.png    \n",
            "  inflating: data/val/6/31942.png    \n",
            "  inflating: data/val/6/31943.png    \n",
            "  inflating: data/val/6/31949.png    \n",
            "  inflating: data/val/6/31955.png    \n",
            "  inflating: data/val/6/31956.png    \n",
            "  inflating: data/val/6/31960.png    \n",
            "  inflating: data/val/6/31963.png    \n",
            "  inflating: data/val/6/31970.png    \n",
            "  inflating: data/val/6/31974.png    \n",
            "  inflating: data/val/6/31981.png    \n",
            "  inflating: data/val/6/31986.png    \n",
            "  inflating: data/val/6/31993.png    \n",
            "  inflating: data/val/6/31994.png    \n",
            "  inflating: data/val/6/31996.png    \n",
            "  inflating: data/val/6/31997.png    \n",
            "  inflating: data/val/6/31999.png    \n",
            "  inflating: data/val/6/32008.png    \n",
            "  inflating: data/val/6/32009.png    \n",
            "  inflating: data/val/6/32016.png    \n",
            "  inflating: data/val/6/32020.png    \n",
            "  inflating: data/val/6/32028.png    \n",
            "  inflating: data/val/6/32032.png    \n",
            "  inflating: data/val/6/32033.png    \n",
            "  inflating: data/val/6/32034.png    \n",
            "  inflating: data/val/6/32049.png    \n",
            "  inflating: data/val/6/32071.png    \n",
            "  inflating: data/val/6/32073.png    \n",
            "  inflating: data/val/6/32081.png    \n",
            "  inflating: data/val/6/32082.png    \n",
            "  inflating: data/val/6/32090.png    \n",
            "  inflating: data/val/6/32101.png    \n",
            "  inflating: data/val/6/32106.png    \n",
            "  inflating: data/val/6/32113.png    \n",
            "  inflating: data/val/6/32115.png    \n",
            "  inflating: data/val/6/32116.png    \n",
            "  inflating: data/val/6/32117.png    \n",
            "  inflating: data/val/6/32118.png    \n",
            "  inflating: data/val/6/32122.png    \n",
            "  inflating: data/val/6/32123.png    \n",
            "  inflating: data/val/6/32127.png    \n",
            "  inflating: data/val/6/32133.png    \n",
            "  inflating: data/val/6/32149.png    \n",
            "  inflating: data/val/6/32155.png    \n",
            "  inflating: data/val/6/32158.png    \n",
            "  inflating: data/val/6/32187.png    \n",
            "  inflating: data/val/6/32191.png    \n",
            "  inflating: data/val/6/32192.png    \n",
            "  inflating: data/val/6/32213.png    \n",
            "  inflating: data/val/6/32216.png    \n",
            "  inflating: data/val/6/32219.png    \n",
            "  inflating: data/val/6/32220.png    \n",
            "  inflating: data/val/6/32222.png    \n",
            "  inflating: data/val/6/32228.png    \n",
            "  inflating: data/val/6/32230.png    \n",
            "  inflating: data/val/6/32239.png    \n",
            "  inflating: data/val/6/32242.png    \n",
            "  inflating: data/val/6/32253.png    \n",
            "  inflating: data/val/6/32269.png    \n",
            "  inflating: data/val/6/32270.png    \n",
            "  inflating: data/val/6/32272.png    \n",
            "  inflating: data/val/6/32285.png    \n"
          ]
        }
      ],
      "source": [
        "!unzip facial-expression-dataset-image-folders-fer2013.zip"
      ],
      "id": "Z_tiH6quKu4-"
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "54349ec1"
      },
      "outputs": [],
      "source": [
        "import os\n",
        "import itertools # To join a list of lists\n",
        "import matplotlib.pyplot as plt # For general plotting\n",
        "import numpy as np\n",
        "from sklearn.metrics import confusion_matrix, ConfusionMatrixDisplay\n",
        "from sklearn.model_selection import train_test_split\n",
        "import torch\n",
        "import torch.nn as nn\n",
        "import torch.nn.functional as F\n",
        "from torch.utils.data import DataLoader, Subset\n",
        "import torchvision\n",
        "import torchvision.transforms as transforms\n",
        "\n",
        "\n",
        "# Utility to visualize PyTorch network and shapes\n",
        "from torchsummary import summary\n",
        "np.set_printoptions(suppress=True)\n",
        "\n",
        "# Set seed to generate reproducible \"pseudo-randomness\" (handles scipy's \"randomness\" too)\n",
        "random_seed = 7\n",
        "np.random.seed(random_seed)\n",
        "torch.manual_seed(7)\n",
        "\n",
        "plt.rc('font', size=22)          # controls default text sizes\n",
        "plt.rc('axes', titlesize=18)     # fontsize of the axes title\n",
        "plt.rc('axes', labelsize=18)     # fontsize of the x and y labels\n",
        "plt.rc('xtick', labelsize=14)    # fontsize of the tick labels\n",
        "plt.rc('ytick', labelsize=14)    # fontsize of the tick labels\n",
        "plt.rc('legend', fontsize=16)    # legend fontsize\n",
        "plt.rc('figure', titlesize=22)   # fontsize of the figure title"
      ],
      "id": "54349ec1"
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 332
        },
        "id": "6ec1401b",
        "outputId": "dbf7f546-3cbd-4a0f-ed55-3308f6c8ee34"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Train dataset has N = 28709 samples\n",
            "Test dataset has N = 3589 samples\n",
            "Validation dataset has N = 3589 samples\n",
            "\n",
            "Labels corresponding to randomly drawn images from the training set:\n",
            "surprise\t happy\t sad  \t surprise\t sad  \t surprise\t angry\t happy\t\n"
          ]
        },
        {
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 1440x1440 with 1 Axes>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        }
      ],
      "source": [
        "device = torch.device(\"cuda:0\" if torch.cuda.is_available() else \"cpu\")\n",
        "\n",
        "# Define how transform the data for processing in batches\n",
        "transform = transforms.Compose([\n",
        "        transforms.Resize([280,280]), # Resizing the image\n",
        "        transforms.ToTensor(),\n",
        "        transforms.Normalize(mean=(0.5,0.5,0.5), std=(0.5,0.5,0.5)) # 0.5 due to how torchvision data is in range [0,1]\n",
        "    ])\n",
        "# Specify the batch size for SGD\n",
        "batch_size = 64\n",
        "\n",
        "# Use the ImageFolder class, which is a generic dataloder useful when images stored in labelled directories\n",
        "# Apply transform to images specified above, importantly converting \"ToTensor\"\n",
        "train_dataset = torchvision.datasets.ImageFolder(root='data/train', transform=transform)\n",
        "test_dataset = torchvision.datasets.ImageFolder(root='data/test', transform=transform)\n",
        "valid_dataset = torchvision.datasets.ImageFolder(root='data/val', transform=transform)\n",
        "\n",
        "classes = ('angry', 'disgust', 'fear', 'happy', 'sad', 'surprise', 'neutral')\n",
        "print(\"Train dataset has N = {} samples\".format(len(train_dataset)))\n",
        "print(\"Test dataset has N = {} samples\".format(len(test_dataset)))\n",
        "print(\"Validation dataset has N = {} samples\".format(len(valid_dataset)))\n",
        "\n",
        "# Define iterable for our newly created dataset and shuffle samples \n",
        "train_dataloader = DataLoader(train_dataset, batch_size=batch_size, shuffle=True)\n",
        "test_dataloader = DataLoader(test_dataset, batch_size=batch_size, shuffle=True)\n",
        "\n",
        "### Plotting stuff for visualization ###\n",
        "fig = plt.figure(figsize=(20,20))\n",
        "# Utility function to show an image\n",
        "def imshow(img):\n",
        "    img = img / 2 + 0.5     # unnormalize\n",
        "    img_np = img.cpu().numpy()\n",
        "    plt.imshow(np.transpose(img_np, (1, 2, 0)))\n",
        "\n",
        "\n",
        "# Get some random training dataset images\n",
        "dataiter = iter(train_dataloader)\n",
        "# Extract a batch\n",
        "images, labels = dataiter.next()\n",
        "images, labels = images.to(device), labels.to(device)\n",
        "\n",
        "# Show 8 images for display\n",
        "num_display = 8\n",
        "imshow(torchvision.utils.make_grid(images[:num_display]))\n",
        "# Print labels as a concatenated string\n",
        "print(\"\\nLabels corresponding to randomly drawn images from the training set:\")\n",
        "print(\" \".join(f'{classes[labels[j]]:5s}\\t' for j in range(num_display)))"
      ],
      "id": "6ec1401b"
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "56a28e44",
        "outputId": "880eddff-a484-43cb-b959-22cae002bb80"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "----------------------------------------------------------------\n",
            "        Layer (type)               Output Shape         Param #\n",
            "================================================================\n",
            "            Conv2d-1         [-1, 16, 280, 280]             448\n",
            "         MaxPool2d-2         [-1, 16, 140, 140]               0\n",
            "            Conv2d-3         [-1, 32, 140, 140]           4,640\n",
            "       BatchNorm2d-4         [-1, 32, 140, 140]              64\n",
            "         MaxPool2d-5           [-1, 32, 70, 70]               0\n",
            "           Dropout-6           [-1, 32, 70, 70]               0\n",
            "            Conv2d-7           [-1, 64, 70, 70]          18,496\n",
            "       BatchNorm2d-8           [-1, 64, 70, 70]             128\n",
            "         MaxPool2d-9           [-1, 64, 35, 35]               0\n",
            "          Dropout-10           [-1, 64, 35, 35]               0\n",
            "           Conv2d-11          [-1, 128, 35, 35]          73,856\n",
            "      BatchNorm2d-12          [-1, 128, 35, 35]             256\n",
            "        MaxPool2d-13          [-1, 128, 17, 17]               0\n",
            "          Dropout-14          [-1, 128, 17, 17]               0\n",
            "           Conv2d-15          [-1, 256, 17, 17]         295,168\n",
            "      BatchNorm2d-16          [-1, 256, 17, 17]             512\n",
            "        MaxPool2d-17            [-1, 256, 8, 8]               0\n",
            "          Dropout-18            [-1, 256, 8, 8]               0\n",
            "           Linear-19                   [-1, 16]         262,160\n",
            "          Dropout-20                   [-1, 16]               0\n",
            "      BatchNorm1d-21                   [-1, 16]              32\n",
            "           Linear-22                   [-1, 16]             272\n",
            "          Dropout-23                   [-1, 16]               0\n",
            "      BatchNorm1d-24                   [-1, 16]              32\n",
            "           Linear-25                    [-1, 7]             119\n",
            "================================================================\n",
            "Total params: 656,183\n",
            "Trainable params: 656,183\n",
            "Non-trainable params: 0\n",
            "----------------------------------------------------------------\n",
            "Input size (MB): 0.90\n",
            "Forward/backward pass size (MB): 34.24\n",
            "Params size (MB): 2.50\n",
            "Estimated Total Size (MB): 37.64\n",
            "----------------------------------------------------------------\n"
          ]
        }
      ],
      "source": [
        "# Note that PyTorch uses NCHW (samples, channels, height, width) image format convention\n",
        "class ConvNet(nn.Module):\n",
        "    def __init__(self, in_channels, num_filters, out_classes, kernel_size=3):\n",
        "        super().__init__()\n",
        "        # Conv2D layer with 'same' padding so image retains shape\n",
        "        self.conv1 = nn.Conv2d(in_channels, num_filters, kernel_size, padding='same')\n",
        "        # self.conv2_bn = nn.BatchNorm2d(num_filters)\n",
        "        self.drop = nn.Dropout(p=0.2)\n",
        "        self.drop2 =nn.Dropout(p=0.2)\n",
        "        #self.dropout = \n",
        "        # MaxPooling layer with 2x2 kernel size and stride 2\n",
        "        self.pool = nn.MaxPool2d(2, 2)\n",
        "        self.conv2 = nn.Conv2d(num_filters, num_filters*2, kernel_size, padding='same')\n",
        "        self.conv2_bn = nn.BatchNorm2d(num_filters*2)\n",
        "        self.conv3 = nn.Conv2d(num_filters*2, num_filters*4, kernel_size, padding='same')\n",
        "        self.conv3_bn = nn.BatchNorm2d(num_filters*4)\n",
        "        self.conv4 = nn.Conv2d(num_filters*4, num_filters*8, kernel_size, padding='same')\n",
        "        self.conv4_bn = nn.BatchNorm2d(num_filters*8)\n",
        "        self.conv5 = nn.Conv2d(num_filters*8, num_filters*16, kernel_size, padding='same')\n",
        "        self.conv5_bn = nn.BatchNorm2d(num_filters*16)\n",
        "        self.fc1 = nn.Linear(num_filters*16*8*8, num_filters)\n",
        "        self.fc1_norm= nn.BatchNorm1d(num_filters)\n",
        "        self.fc2 = nn.Linear(num_filters, num_filters)\n",
        "        self.fc2_norm= nn.BatchNorm1d(num_filters)\n",
        "        self.fc3 = nn.Linear(num_filters, out_classes)\n",
        "\n",
        "\n",
        "    def forward(self, x):\n",
        "        # Non-linear ReLU activations between convolutional layers\n",
        "        # Conv->ReLU->Pooling\n",
        "        # 280x280 image -> 140x140 after pooling\n",
        "        x = self.pool(F.relu(self.conv1(x)))\n",
        "        # print(x.shape)\n",
        "        # 140x140 feature map -> 70x70 after pooling\n",
        "        x = self.drop(self.pool(self.conv2_bn(F.relu(self.conv2(x)))))\n",
        "        # print(x.shape)\n",
        "        # Flatten all dimensions except batch (start_dim=1)\n",
        "        x = self.drop(self.pool(self.conv3_bn(F.relu(self.conv3(x)))))\n",
        "        # print(x.shape)\n",
        "        x = self.drop(self.pool(self.conv4_bn(F.relu(self.conv4(x)))))\n",
        "        x = self.drop(self.pool(self.conv5_bn(F.relu(self.conv5(x)))))\n",
        "        x = torch.flatten(x, 1) \n",
        "        # print(x.shape)\n",
        "        x = self.fc1_norm(F.relu(self.drop2(self.fc1(x))))\n",
        "        x = self.fc2_norm(F.relu(self.drop2(self.fc2(x))))\n",
        "        x = self.fc3(x)\n",
        "        # print(\"shape is: \", self.conv1.size, self.conv2.size, self.conv3.size, self.fc1.size, self.fc2.size)\n",
        "        \n",
        "        return x\n",
        "\n",
        "input_channels = images.shape[1]\n",
        "model = ConvNet(input_channels, num_filters=16, out_classes=len(classes))\n",
        "\n",
        "# Dummy inputs so we can plot a summary of the neural network's architecture and no. of parameters\n",
        "model=model.to(device)\n",
        "summary(model, input_size=(input_channels, 280, 280))\n",
        "\n",
        "# Double check if model exists\n",
        "# if os.path.exists(os.getcwd() + '/cnn_fmd.pth'):\n",
        "#     model.load_state_dict(torch.load(os.getcwd() + '/cnn_fmd.pth'))\n",
        "#     print(\"Loaded model from disk!\")"
      ],
      "id": "56a28e44"
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "background_save": true,
          "base_uri": "https://localhost:8080/"
        },
        "id": "DTe-UYFiV2vo",
        "outputId": "4a6af11d-06d3-43a7-d22e-4b0a3f78ddf9"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Epoch 1\n",
            "-------------------------------\n",
            "loss: 2.177123  [    0/28709]\n",
            "loss: 2.296377  [  640/28709]\n",
            "loss: 2.084889  [ 1280/28709]\n",
            "loss: 1.916442  [ 1920/28709]\n",
            "loss: 2.148739  [ 2560/28709]\n",
            "loss: 2.049479  [ 3200/28709]\n",
            "loss: 2.038434  [ 3840/28709]\n",
            "loss: 2.070794  [ 4480/28709]\n",
            "loss: 1.926384  [ 5120/28709]\n",
            "loss: 2.048480  [ 5760/28709]\n",
            "loss: 2.075200  [ 6400/28709]\n",
            "loss: 2.042711  [ 7040/28709]\n",
            "loss: 2.055386  [ 7680/28709]\n",
            "loss: 1.927482  [ 8320/28709]\n",
            "loss: 1.999658  [ 8960/28709]\n",
            "loss: 2.077408  [ 9600/28709]\n",
            "loss: 2.004540  [10240/28709]\n",
            "loss: 2.007305  [10880/28709]\n",
            "loss: 1.945045  [11520/28709]\n",
            "loss: 1.869569  [12160/28709]\n",
            "loss: 1.938420  [12800/28709]\n",
            "loss: 2.056867  [13440/28709]\n",
            "loss: 1.944896  [14080/28709]\n",
            "loss: 1.905140  [14720/28709]\n",
            "loss: 1.958518  [15360/28709]\n",
            "loss: 1.858966  [16000/28709]\n",
            "loss: 1.992688  [16640/28709]\n",
            "loss: 1.791454  [17280/28709]\n",
            "loss: 1.922023  [17920/28709]\n",
            "loss: 2.033635  [18560/28709]\n",
            "loss: 1.912454  [19200/28709]\n",
            "loss: 1.852150  [19840/28709]\n",
            "loss: 1.871082  [20480/28709]\n",
            "loss: 1.869769  [21120/28709]\n",
            "loss: 1.968224  [21760/28709]\n",
            "loss: 1.878806  [22400/28709]\n",
            "loss: 1.778875  [23040/28709]\n",
            "loss: 1.820660  [23680/28709]\n",
            "loss: 1.934681  [24320/28709]\n",
            "loss: 1.816966  [24960/28709]\n",
            "loss: 1.881407  [25600/28709]\n",
            "loss: 1.900235  [26240/28709]\n",
            "loss: 1.926482  [26880/28709]\n",
            "loss: 1.926642  [27520/28709]\n",
            "loss: 1.867159  [28160/28709]\n",
            "\n",
            "Test Error\n",
            "\tAccuracy: 31.4%, Avg loss: 2.183611 \n",
            "\n",
            "Epoch 2\n",
            "-------------------------------\n",
            "loss: 1.926659  [    0/28709]\n",
            "loss: 1.926641  [  640/28709]\n",
            "loss: 1.761183  [ 1280/28709]\n",
            "loss: 1.815393  [ 1920/28709]\n",
            "loss: 1.806023  [ 2560/28709]\n",
            "loss: 1.857975  [ 3200/28709]\n",
            "loss: 1.835752  [ 3840/28709]\n",
            "loss: 1.839055  [ 4480/28709]\n",
            "loss: 1.856701  [ 5120/28709]\n",
            "loss: 1.867188  [ 5760/28709]\n",
            "loss: 1.763504  [ 6400/28709]\n",
            "loss: 1.932585  [ 7040/28709]\n",
            "loss: 1.745829  [ 7680/28709]\n",
            "loss: 1.814988  [ 8320/28709]\n",
            "loss: 1.859974  [ 8960/28709]\n",
            "loss: 1.776196  [ 9600/28709]\n",
            "loss: 1.765879  [10240/28709]\n",
            "loss: 1.782739  [10880/28709]\n",
            "loss: 1.781857  [11520/28709]\n",
            "loss: 1.810941  [12160/28709]\n",
            "loss: 1.736646  [12800/28709]\n",
            "loss: 1.838747  [13440/28709]\n",
            "loss: 1.800775  [14080/28709]\n",
            "loss: 1.813655  [14720/28709]\n",
            "loss: 1.748196  [15360/28709]\n",
            "loss: 1.704839  [16000/28709]\n",
            "loss: 1.897541  [16640/28709]\n",
            "loss: 1.784767  [17280/28709]\n",
            "loss: 1.728441  [17920/28709]\n",
            "loss: 1.709823  [18560/28709]\n",
            "loss: 1.785801  [19200/28709]\n",
            "loss: 1.748922  [19840/28709]\n",
            "loss: 1.761286  [20480/28709]\n",
            "loss: 1.786790  [21120/28709]\n",
            "loss: 1.673705  [21760/28709]\n",
            "loss: 1.815511  [22400/28709]\n",
            "loss: 1.783878  [23040/28709]\n",
            "loss: 1.747015  [23680/28709]\n",
            "loss: 1.593988  [24320/28709]\n",
            "loss: 1.630449  [24960/28709]\n",
            "loss: 1.735369  [25600/28709]\n",
            "loss: 1.763180  [26240/28709]\n",
            "loss: 1.879650  [26880/28709]\n",
            "loss: 1.854108  [27520/28709]\n",
            "loss: 1.734968  [28160/28709]\n",
            "\n",
            "Test Error\n",
            "\tAccuracy: 35.6%, Avg loss: 1.730475 \n",
            "\n",
            "Epoch 3\n",
            "-------------------------------\n",
            "loss: 1.692915  [    0/28709]\n",
            "loss: 1.777308  [  640/28709]\n",
            "loss: 1.745442  [ 1280/28709]\n",
            "loss: 1.735001  [ 1920/28709]\n",
            "loss: 1.604403  [ 2560/28709]\n",
            "loss: 1.676491  [ 3200/28709]\n",
            "loss: 1.702011  [ 3840/28709]\n",
            "loss: 1.743628  [ 4480/28709]\n",
            "loss: 1.809413  [ 5120/28709]\n",
            "loss: 1.782109  [ 5760/28709]\n",
            "loss: 1.850440  [ 6400/28709]\n",
            "loss: 1.675652  [ 7040/28709]\n",
            "loss: 1.771553  [ 7680/28709]\n",
            "loss: 1.761749  [ 8320/28709]\n",
            "loss: 1.729252  [ 8960/28709]\n",
            "loss: 1.763095  [ 9600/28709]\n",
            "loss: 1.643508  [10240/28709]\n",
            "loss: 1.810490  [10880/28709]\n",
            "loss: 1.720603  [11520/28709]\n",
            "loss: 1.830598  [12160/28709]\n",
            "loss: 1.850921  [12800/28709]\n",
            "loss: 1.742411  [13440/28709]\n",
            "loss: 1.687424  [14080/28709]\n",
            "loss: 1.657969  [14720/28709]\n",
            "loss: 1.691797  [15360/28709]\n",
            "loss: 1.683943  [16000/28709]\n",
            "loss: 2.032067  [16640/28709]\n",
            "loss: 1.619671  [17280/28709]\n",
            "loss: 1.689671  [17920/28709]\n",
            "loss: 1.681019  [18560/28709]\n",
            "loss: 1.730910  [19200/28709]\n",
            "loss: 1.685318  [19840/28709]\n",
            "loss: 1.759278  [20480/28709]\n",
            "loss: 1.634173  [21120/28709]\n",
            "loss: 1.651317  [21760/28709]\n",
            "loss: 1.558909  [22400/28709]\n",
            "loss: 1.815328  [23040/28709]\n",
            "loss: 1.679333  [23680/28709]\n",
            "loss: 1.670286  [24320/28709]\n",
            "loss: 1.563262  [24960/28709]\n",
            "loss: 1.738199  [25600/28709]\n",
            "loss: 1.572553  [26240/28709]\n",
            "loss: 1.635037  [26880/28709]\n",
            "loss: 1.592693  [27520/28709]\n",
            "loss: 1.583755  [28160/28709]\n",
            "\n",
            "Test Error\n",
            "\tAccuracy: 41.5%, Avg loss: 1.589835 \n",
            "\n",
            "Epoch 4\n",
            "-------------------------------\n",
            "loss: 1.771782  [    0/28709]\n",
            "loss: 1.646198  [  640/28709]\n",
            "loss: 1.572709  [ 1280/28709]\n",
            "loss: 1.560169  [ 1920/28709]\n",
            "loss: 1.760683  [ 2560/28709]\n",
            "loss: 1.584435  [ 3200/28709]\n",
            "loss: 1.656777  [ 3840/28709]\n",
            "loss: 1.652663  [ 4480/28709]\n",
            "loss: 1.540686  [ 5120/28709]\n",
            "loss: 1.669363  [ 5760/28709]\n",
            "loss: 1.647605  [ 6400/28709]\n",
            "loss: 1.903579  [ 7040/28709]\n",
            "loss: 1.556522  [ 7680/28709]\n",
            "loss: 1.591194  [ 8320/28709]\n",
            "loss: 1.703904  [ 8960/28709]\n",
            "loss: 1.605155  [ 9600/28709]\n",
            "loss: 1.485414  [10240/28709]\n",
            "loss: 1.562542  [10880/28709]\n",
            "loss: 1.554847  [11520/28709]\n",
            "loss: 1.633988  [12160/28709]\n",
            "loss: 1.661574  [12800/28709]\n",
            "loss: 1.536432  [13440/28709]\n",
            "loss: 1.654138  [14080/28709]\n",
            "loss: 1.619574  [14720/28709]\n",
            "loss: 1.753140  [15360/28709]\n",
            "loss: 1.708159  [16000/28709]\n",
            "loss: 1.768867  [16640/28709]\n",
            "loss: 1.607354  [17280/28709]\n",
            "loss: 1.611738  [17920/28709]\n",
            "loss: 1.634199  [18560/28709]\n",
            "loss: 1.454578  [19200/28709]\n",
            "loss: 1.711391  [19840/28709]\n",
            "loss: 1.513772  [20480/28709]\n",
            "loss: 1.624169  [21120/28709]\n",
            "loss: 1.545220  [21760/28709]\n",
            "loss: 1.479443  [22400/28709]\n",
            "loss: 1.634745  [23040/28709]\n",
            "loss: 1.635201  [23680/28709]\n",
            "loss: 1.633075  [24320/28709]\n",
            "loss: 1.645740  [24960/28709]\n",
            "loss: 1.672329  [25600/28709]\n",
            "loss: 1.572915  [26240/28709]\n",
            "loss: 1.621073  [26880/28709]\n",
            "loss: 1.687883  [27520/28709]\n",
            "loss: 1.484998  [28160/28709]\n",
            "\n",
            "Test Error\n",
            "\tAccuracy: 44.5%, Avg loss: 1.517408 \n",
            "\n",
            "Epoch 5\n",
            "-------------------------------\n",
            "loss: 1.580522  [    0/28709]\n",
            "loss: 1.462694  [  640/28709]\n",
            "loss: 1.678809  [ 1280/28709]\n",
            "loss: 1.566246  [ 1920/28709]\n",
            "loss: 1.481228  [ 2560/28709]\n",
            "loss: 1.706796  [ 3200/28709]\n",
            "loss: 1.449830  [ 3840/28709]\n",
            "loss: 1.550599  [ 4480/28709]\n",
            "loss: 1.547125  [ 5120/28709]\n",
            "loss: 1.637523  [ 5760/28709]\n",
            "loss: 1.501475  [ 6400/28709]\n",
            "loss: 1.479586  [ 7040/28709]\n",
            "loss: 1.538142  [ 7680/28709]\n",
            "loss: 1.537941  [ 8320/28709]\n",
            "loss: 1.597039  [ 8960/28709]\n",
            "loss: 1.643521  [ 9600/28709]\n",
            "loss: 1.638130  [10240/28709]\n",
            "loss: 1.627654  [10880/28709]\n",
            "loss: 1.513751  [11520/28709]\n",
            "loss: 1.586587  [12160/28709]\n",
            "loss: 1.539613  [12800/28709]\n",
            "loss: 1.514683  [13440/28709]\n",
            "loss: 1.505582  [14080/28709]\n",
            "loss: 1.433746  [14720/28709]\n",
            "loss: 1.567438  [15360/28709]\n",
            "loss: 1.599553  [16000/28709]\n",
            "loss: 1.603987  [16640/28709]\n",
            "loss: 1.476373  [17280/28709]\n",
            "loss: 1.604363  [17920/28709]\n",
            "loss: 1.528770  [18560/28709]\n",
            "loss: 1.614066  [19200/28709]\n",
            "loss: 1.551546  [19840/28709]\n",
            "loss: 1.704931  [20480/28709]\n",
            "loss: 1.427214  [21120/28709]\n",
            "loss: 1.527592  [21760/28709]\n",
            "loss: 1.405314  [22400/28709]\n",
            "loss: 1.637077  [23040/28709]\n",
            "loss: 1.674595  [23680/28709]\n",
            "loss: 1.553393  [24320/28709]\n",
            "loss: 1.646650  [24960/28709]\n",
            "loss: 1.558794  [25600/28709]\n",
            "loss: 1.533750  [26240/28709]\n",
            "loss: 1.395463  [26880/28709]\n",
            "loss: 1.520702  [27520/28709]\n",
            "loss: 1.597379  [28160/28709]\n",
            "\n",
            "Test Error\n",
            "\tAccuracy: 45.7%, Avg loss: 1.470252 \n",
            "\n",
            "Epoch 6\n",
            "-------------------------------\n",
            "loss: 1.472119  [    0/28709]\n",
            "loss: 1.472610  [  640/28709]\n",
            "loss: 1.510438  [ 1280/28709]\n",
            "loss: 1.448925  [ 1920/28709]\n",
            "loss: 1.568648  [ 2560/28709]\n",
            "loss: 1.571652  [ 3200/28709]\n",
            "loss: 1.397697  [ 3840/28709]\n",
            "loss: 1.449999  [ 4480/28709]\n",
            "loss: 1.527214  [ 5120/28709]\n",
            "loss: 1.464679  [ 5760/28709]\n",
            "loss: 1.351377  [ 6400/28709]\n",
            "loss: 1.693358  [ 7040/28709]\n",
            "loss: 1.401307  [ 7680/28709]\n",
            "loss: 1.463873  [ 8320/28709]\n",
            "loss: 1.427691  [ 8960/28709]\n",
            "loss: 1.472787  [ 9600/28709]\n",
            "loss: 1.437270  [10240/28709]\n",
            "loss: 1.372632  [10880/28709]\n",
            "loss: 1.556715  [11520/28709]\n",
            "loss: 1.617764  [12160/28709]\n",
            "loss: 1.404829  [12800/28709]\n",
            "loss: 1.585464  [13440/28709]\n",
            "loss: 1.641049  [14080/28709]\n",
            "loss: 1.515209  [14720/28709]\n",
            "loss: 1.428827  [15360/28709]\n",
            "loss: 1.505437  [16000/28709]\n",
            "loss: 1.635641  [16640/28709]\n",
            "loss: 1.471849  [17280/28709]\n",
            "loss: 1.602873  [17920/28709]\n",
            "loss: 1.591806  [18560/28709]\n",
            "loss: 1.569084  [19200/28709]\n",
            "loss: 1.511748  [19840/28709]\n",
            "loss: 1.540239  [20480/28709]\n",
            "loss: 1.378620  [21120/28709]\n",
            "loss: 1.402402  [21760/28709]\n",
            "loss: 1.424215  [22400/28709]\n",
            "loss: 1.478888  [23040/28709]\n",
            "loss: 1.415462  [23680/28709]\n",
            "loss: 1.573882  [24320/28709]\n",
            "loss: 1.546295  [24960/28709]\n",
            "loss: 1.567674  [25600/28709]\n",
            "loss: 1.411707  [26240/28709]\n",
            "loss: 1.579231  [26880/28709]\n",
            "loss: 1.488893  [27520/28709]\n",
            "loss: 1.660305  [28160/28709]\n",
            "\n",
            "Test Error\n",
            "\tAccuracy: 45.8%, Avg loss: 1.434961 \n",
            "\n",
            "Epoch 7\n",
            "-------------------------------\n",
            "loss: 1.390275  [    0/28709]\n",
            "loss: 1.588300  [  640/28709]\n",
            "loss: 1.397738  [ 1280/28709]\n",
            "loss: 1.573891  [ 1920/28709]\n",
            "loss: 1.547385  [ 2560/28709]\n",
            "loss: 1.463059  [ 3200/28709]\n",
            "loss: 1.605707  [ 3840/28709]\n",
            "loss: 1.479640  [ 4480/28709]\n",
            "loss: 1.439905  [ 5120/28709]\n",
            "loss: 1.383020  [ 5760/28709]\n",
            "loss: 1.544448  [ 6400/28709]\n",
            "loss: 1.553864  [ 7040/28709]\n",
            "loss: 1.653966  [ 7680/28709]\n",
            "loss: 1.534731  [ 8320/28709]\n",
            "loss: 1.489731  [ 8960/28709]\n",
            "loss: 1.512224  [ 9600/28709]\n",
            "loss: 1.345384  [10240/28709]\n",
            "loss: 1.567899  [10880/28709]\n",
            "loss: 1.361682  [11520/28709]\n",
            "loss: 1.515764  [12160/28709]\n",
            "loss: 1.671690  [12800/28709]\n",
            "loss: 1.455525  [13440/28709]\n",
            "loss: 1.287192  [14080/28709]\n",
            "loss: 1.397320  [14720/28709]\n",
            "loss: 1.580308  [15360/28709]\n",
            "loss: 1.828667  [16000/28709]\n",
            "loss: 1.560246  [16640/28709]\n",
            "loss: 1.373297  [17280/28709]\n",
            "loss: 1.436598  [17920/28709]\n",
            "loss: 1.370198  [18560/28709]\n",
            "loss: 1.348429  [19200/28709]\n",
            "loss: 1.302046  [19840/28709]\n",
            "loss: 1.335600  [20480/28709]\n",
            "loss: 1.456996  [21120/28709]\n",
            "loss: 1.501238  [21760/28709]\n",
            "loss: 1.449450  [22400/28709]\n",
            "loss: 1.563647  [23040/28709]\n",
            "loss: 1.533994  [23680/28709]\n",
            "loss: 1.476439  [24320/28709]\n",
            "loss: 1.468858  [24960/28709]\n",
            "loss: 1.433407  [25600/28709]\n",
            "loss: 1.479182  [26240/28709]\n",
            "loss: 1.469620  [26880/28709]\n",
            "loss: 1.448921  [27520/28709]\n",
            "loss: 1.540502  [28160/28709]\n",
            "\n",
            "Test Error\n",
            "\tAccuracy: 48.7%, Avg loss: 1.388107 \n",
            "\n",
            "Epoch 8\n",
            "-------------------------------\n",
            "loss: 1.375464  [    0/28709]\n",
            "loss: 1.301305  [  640/28709]\n",
            "loss: 1.476469  [ 1280/28709]\n",
            "loss: 1.494405  [ 1920/28709]\n",
            "loss: 1.404348  [ 2560/28709]\n",
            "loss: 1.493964  [ 3200/28709]\n",
            "loss: 1.333228  [ 3840/28709]\n",
            "loss: 1.572150  [ 4480/28709]\n",
            "loss: 1.271179  [ 5120/28709]\n",
            "loss: 1.313903  [ 5760/28709]\n",
            "loss: 1.386543  [ 6400/28709]\n",
            "loss: 1.400313  [ 7040/28709]\n",
            "loss: 1.444347  [ 7680/28709]\n",
            "loss: 1.404232  [ 8320/28709]\n",
            "loss: 1.580230  [ 8960/28709]\n",
            "loss: 1.498075  [ 9600/28709]\n",
            "loss: 1.365436  [10240/28709]\n",
            "loss: 1.530939  [10880/28709]\n",
            "loss: 1.365642  [11520/28709]\n",
            "loss: 1.347298  [12160/28709]\n",
            "loss: 1.525675  [12800/28709]\n",
            "loss: 1.757244  [13440/28709]\n",
            "loss: 1.462672  [14080/28709]\n",
            "loss: 1.343362  [14720/28709]\n",
            "loss: 1.436283  [15360/28709]\n",
            "loss: 1.370663  [16000/28709]\n",
            "loss: 1.542113  [16640/28709]\n",
            "loss: 1.547552  [17280/28709]\n",
            "loss: 1.493162  [17920/28709]\n",
            "loss: 1.478984  [18560/28709]\n",
            "loss: 1.258921  [19200/28709]\n",
            "loss: 1.275032  [19840/28709]\n",
            "loss: 1.441781  [20480/28709]\n",
            "loss: 1.541677  [21120/28709]\n",
            "loss: 1.407327  [21760/28709]\n",
            "loss: 1.360397  [22400/28709]\n",
            "loss: 1.323607  [23040/28709]\n",
            "loss: 1.485368  [23680/28709]\n",
            "loss: 1.599015  [24320/28709]\n",
            "loss: 1.423691  [24960/28709]\n",
            "loss: 1.302724  [25600/28709]\n",
            "loss: 1.387765  [26240/28709]\n",
            "loss: 1.494558  [26880/28709]\n",
            "loss: 1.603121  [27520/28709]\n",
            "loss: 1.343040  [28160/28709]\n",
            "\n",
            "Test Error\n",
            "\tAccuracy: 48.5%, Avg loss: 1.404052 \n",
            "\n",
            "Epoch 9\n",
            "-------------------------------\n",
            "loss: 1.455416  [    0/28709]\n",
            "loss: 1.428599  [  640/28709]\n",
            "loss: 1.422757  [ 1280/28709]\n",
            "loss: 1.353039  [ 1920/28709]\n",
            "loss: 1.379062  [ 2560/28709]\n",
            "loss: 1.315641  [ 3200/28709]\n",
            "loss: 1.442067  [ 3840/28709]\n",
            "loss: 1.553016  [ 4480/28709]\n",
            "loss: 1.332596  [ 5120/28709]\n",
            "loss: 1.449257  [ 5760/28709]\n",
            "loss: 1.355982  [ 6400/28709]\n",
            "loss: 1.355518  [ 7040/28709]\n",
            "loss: 1.461225  [ 7680/28709]\n",
            "loss: 1.420821  [ 8320/28709]\n",
            "loss: 1.708358  [ 8960/28709]\n",
            "loss: 1.291746  [ 9600/28709]\n",
            "loss: 1.313548  [10240/28709]\n",
            "loss: 1.295771  [10880/28709]\n",
            "loss: 1.402891  [11520/28709]\n",
            "loss: 1.312122  [12160/28709]\n",
            "loss: 1.590797  [12800/28709]\n",
            "loss: 1.466239  [13440/28709]\n",
            "loss: 1.401231  [14080/28709]\n",
            "loss: 1.440525  [14720/28709]\n",
            "loss: 1.487017  [15360/28709]\n",
            "loss: 1.270144  [16000/28709]\n",
            "loss: 1.324972  [16640/28709]\n",
            "loss: 1.557677  [17280/28709]\n",
            "loss: 1.662093  [17920/28709]\n",
            "loss: 1.654713  [18560/28709]\n",
            "loss: 1.399882  [19200/28709]\n",
            "loss: 1.499232  [19840/28709]\n",
            "loss: 1.279449  [20480/28709]\n",
            "loss: 1.358864  [21120/28709]\n",
            "loss: 1.251989  [21760/28709]\n",
            "loss: 1.353543  [22400/28709]\n",
            "loss: 1.479387  [23040/28709]\n",
            "loss: 1.406921  [23680/28709]\n",
            "loss: 1.422061  [24320/28709]\n",
            "loss: 1.360603  [24960/28709]\n",
            "loss: 1.315348  [25600/28709]\n",
            "loss: 1.466619  [26240/28709]\n",
            "loss: 1.496368  [26880/28709]\n",
            "loss: 1.461892  [27520/28709]\n",
            "loss: 1.672739  [28160/28709]\n",
            "\n",
            "Test Error\n",
            "\tAccuracy: 50.1%, Avg loss: 1.335460 \n",
            "\n",
            "Epoch 10\n",
            "-------------------------------\n",
            "loss: 1.221393  [    0/28709]\n",
            "loss: 1.433135  [  640/28709]\n",
            "loss: 1.290162  [ 1280/28709]\n",
            "loss: 1.357280  [ 1920/28709]\n",
            "loss: 1.329247  [ 2560/28709]\n",
            "loss: 1.379582  [ 3200/28709]\n",
            "loss: 1.249202  [ 3840/28709]\n",
            "loss: 1.298884  [ 4480/28709]\n",
            "loss: 1.526456  [ 5120/28709]\n",
            "loss: 1.286296  [ 5760/28709]\n",
            "loss: 1.236041  [ 6400/28709]\n",
            "loss: 1.438782  [ 7040/28709]\n",
            "loss: 1.273526  [ 7680/28709]\n",
            "loss: 1.316443  [ 8320/28709]\n",
            "loss: 1.242867  [ 8960/28709]\n",
            "loss: 1.369178  [ 9600/28709]\n",
            "loss: 1.632918  [10240/28709]\n",
            "loss: 1.326122  [10880/28709]\n",
            "loss: 1.252860  [11520/28709]\n",
            "loss: 1.355539  [12160/28709]\n",
            "loss: 1.508997  [12800/28709]\n",
            "loss: 1.442417  [13440/28709]\n",
            "loss: 1.468097  [14080/28709]\n",
            "loss: 1.345685  [14720/28709]\n",
            "loss: 1.290110  [15360/28709]\n",
            "loss: 1.474899  [16000/28709]\n",
            "loss: 1.574237  [16640/28709]\n",
            "loss: 1.193499  [17280/28709]\n",
            "loss: 1.311732  [17920/28709]\n",
            "loss: 1.320109  [18560/28709]\n",
            "loss: 1.384609  [19200/28709]\n",
            "loss: 1.415952  [19840/28709]\n",
            "loss: 1.311095  [20480/28709]\n",
            "loss: 1.409076  [21120/28709]\n",
            "loss: 1.319567  [21760/28709]\n",
            "loss: 1.520641  [22400/28709]\n",
            "loss: 1.343141  [23040/28709]\n",
            "loss: 1.273980  [23680/28709]\n",
            "loss: 1.378912  [24320/28709]\n",
            "loss: 1.521680  [24960/28709]\n",
            "loss: 1.305915  [25600/28709]\n",
            "loss: 1.103079  [26240/28709]\n",
            "loss: 1.426795  [26880/28709]\n",
            "loss: 1.247331  [27520/28709]\n",
            "loss: 1.270018  [28160/28709]\n",
            "\n",
            "Test Error\n",
            "\tAccuracy: 51.4%, Avg loss: 1.296613 \n",
            "\n",
            "Epoch 11\n",
            "-------------------------------\n",
            "loss: 1.346313  [    0/28709]\n",
            "loss: 1.266936  [  640/28709]\n",
            "loss: 1.336909  [ 1280/28709]\n",
            "loss: 1.503947  [ 1920/28709]\n",
            "loss: 1.299786  [ 2560/28709]\n",
            "loss: 1.340511  [ 3200/28709]\n",
            "loss: 1.255862  [ 3840/28709]\n",
            "loss: 1.293041  [ 4480/28709]\n",
            "loss: 1.248908  [ 5120/28709]\n",
            "loss: 1.234742  [ 5760/28709]\n",
            "loss: 1.168057  [ 6400/28709]\n",
            "loss: 1.439134  [ 7040/28709]\n",
            "loss: 1.320196  [ 7680/28709]\n",
            "loss: 1.319769  [ 8320/28709]\n",
            "loss: 1.424995  [ 8960/28709]\n",
            "loss: 1.238630  [ 9600/28709]\n",
            "loss: 1.250980  [10240/28709]\n",
            "loss: 1.123798  [10880/28709]\n",
            "loss: 1.413852  [11520/28709]\n",
            "loss: 1.313129  [12160/28709]\n",
            "loss: 1.283944  [12800/28709]\n",
            "loss: 1.398410  [13440/28709]\n",
            "loss: 1.166899  [14080/28709]\n",
            "loss: 1.422505  [14720/28709]\n",
            "loss: 1.308485  [15360/28709]\n",
            "loss: 1.359126  [16000/28709]\n",
            "loss: 1.358962  [16640/28709]\n",
            "loss: 1.162341  [17280/28709]\n",
            "loss: 1.446316  [17920/28709]\n",
            "loss: 1.292680  [18560/28709]\n",
            "loss: 1.451574  [19200/28709]\n",
            "loss: 1.320936  [19840/28709]\n",
            "loss: 1.191344  [20480/28709]\n",
            "loss: 1.394403  [21120/28709]\n",
            "loss: 1.304103  [21760/28709]\n",
            "loss: 1.231378  [22400/28709]\n",
            "loss: 1.408452  [23040/28709]\n",
            "loss: 1.088950  [23680/28709]\n",
            "loss: 1.133204  [24320/28709]\n",
            "loss: 1.378433  [24960/28709]\n",
            "loss: 1.174508  [25600/28709]\n",
            "loss: 1.335970  [26240/28709]\n",
            "loss: 1.283492  [26880/28709]\n",
            "loss: 1.251062  [27520/28709]\n",
            "loss: 1.253596  [28160/28709]\n",
            "\n",
            "Test Error\n",
            "\tAccuracy: 51.4%, Avg loss: 1.278143 \n",
            "\n",
            "Epoch 12\n",
            "-------------------------------\n",
            "loss: 1.412534  [    0/28709]\n",
            "loss: 1.297174  [  640/28709]\n",
            "loss: 1.225119  [ 1280/28709]\n",
            "loss: 1.254851  [ 1920/28709]\n",
            "loss: 1.357071  [ 2560/28709]\n",
            "loss: 1.366288  [ 3200/28709]\n",
            "loss: 1.406342  [ 3840/28709]\n",
            "loss: 1.279218  [ 4480/28709]\n",
            "loss: 1.394723  [ 5120/28709]\n",
            "loss: 1.185836  [ 5760/28709]\n",
            "loss: 1.243351  [ 6400/28709]\n",
            "loss: 1.024173  [ 7040/28709]\n",
            "loss: 1.415727  [ 7680/28709]\n",
            "loss: 1.094510  [ 8320/28709]\n",
            "loss: 1.208271  [ 8960/28709]\n",
            "loss: 1.290916  [ 9600/28709]\n",
            "loss: 1.188026  [10240/28709]\n",
            "loss: 1.243459  [10880/28709]\n",
            "loss: 1.178556  [11520/28709]\n",
            "loss: 1.311499  [12160/28709]\n",
            "loss: 1.272234  [12800/28709]\n",
            "loss: 1.253585  [13440/28709]\n",
            "loss: 1.193966  [14080/28709]\n",
            "loss: 1.293649  [14720/28709]\n",
            "loss: 1.308262  [15360/28709]\n",
            "loss: 1.377534  [16000/28709]\n",
            "loss: 1.267425  [16640/28709]\n",
            "loss: 1.422647  [17280/28709]\n",
            "loss: 1.317551  [17920/28709]\n",
            "loss: 1.376394  [18560/28709]\n",
            "loss: 1.439363  [19200/28709]\n",
            "loss: 1.136874  [19840/28709]\n",
            "loss: 1.190934  [20480/28709]\n",
            "loss: 1.267333  [21120/28709]\n",
            "loss: 1.429327  [21760/28709]\n",
            "loss: 1.144086  [22400/28709]\n",
            "loss: 1.291472  [23040/28709]\n",
            "loss: 1.247392  [23680/28709]\n",
            "loss: 1.241307  [24320/28709]\n",
            "loss: 1.304625  [24960/28709]\n",
            "loss: 1.122315  [25600/28709]\n",
            "loss: 1.025709  [26240/28709]\n",
            "loss: 1.286558  [26880/28709]\n",
            "loss: 1.450436  [27520/28709]\n",
            "loss: 1.256871  [28160/28709]\n",
            "\n",
            "Test Error\n",
            "\tAccuracy: 53.2%, Avg loss: 1.250467 \n",
            "\n",
            "Epoch 13\n",
            "-------------------------------\n",
            "loss: 1.362085  [    0/28709]\n",
            "loss: 1.277847  [  640/28709]\n",
            "loss: 1.174351  [ 1280/28709]\n",
            "loss: 1.214420  [ 1920/28709]\n",
            "loss: 1.258457  [ 2560/28709]\n",
            "loss: 1.032407  [ 3200/28709]\n",
            "loss: 1.293564  [ 3840/28709]\n",
            "loss: 1.299398  [ 4480/28709]\n",
            "loss: 1.287253  [ 5120/28709]\n",
            "loss: 1.199043  [ 5760/28709]\n",
            "loss: 1.154223  [ 6400/28709]\n",
            "loss: 1.363128  [ 7040/28709]\n",
            "loss: 1.210648  [ 7680/28709]\n",
            "loss: 1.096544  [ 8320/28709]\n",
            "loss: 1.269149  [ 8960/28709]\n",
            "loss: 1.090390  [ 9600/28709]\n",
            "loss: 1.140232  [10240/28709]\n",
            "loss: 1.181969  [10880/28709]\n",
            "loss: 1.138221  [11520/28709]\n",
            "loss: 1.119083  [12160/28709]\n",
            "loss: 1.364639  [12800/28709]\n",
            "loss: 1.268947  [13440/28709]\n",
            "loss: 1.266079  [14080/28709]\n",
            "loss: 1.312112  [14720/28709]\n",
            "loss: 1.381003  [15360/28709]\n",
            "loss: 1.178764  [16000/28709]\n",
            "loss: 1.372409  [16640/28709]\n",
            "loss: 1.398029  [17280/28709]\n",
            "loss: 1.177424  [17920/28709]\n",
            "loss: 1.136658  [18560/28709]\n",
            "loss: 1.074347  [19200/28709]\n",
            "loss: 1.271175  [19840/28709]\n",
            "loss: 1.104805  [20480/28709]\n",
            "loss: 1.215026  [21120/28709]\n",
            "loss: 1.240150  [21760/28709]\n",
            "loss: 1.245838  [22400/28709]\n",
            "loss: 1.272831  [23040/28709]\n",
            "loss: 1.174264  [23680/28709]\n",
            "loss: 1.340899  [24320/28709]\n",
            "loss: 1.038093  [24960/28709]\n",
            "loss: 1.004249  [25600/28709]\n",
            "loss: 1.105805  [26240/28709]\n",
            "loss: 1.323655  [26880/28709]\n",
            "loss: 1.107197  [27520/28709]\n",
            "loss: 1.184247  [28160/28709]\n",
            "\n",
            "Test Error\n",
            "\tAccuracy: 53.7%, Avg loss: 1.234582 \n",
            "\n",
            "Epoch 14\n",
            "-------------------------------\n",
            "loss: 1.484325  [    0/28709]\n",
            "loss: 1.287332  [  640/28709]\n",
            "loss: 1.230507  [ 1280/28709]\n",
            "loss: 1.192654  [ 1920/28709]\n",
            "loss: 1.329880  [ 2560/28709]\n",
            "loss: 1.179477  [ 3200/28709]\n",
            "loss: 1.157078  [ 3840/28709]\n",
            "loss: 1.297891  [ 4480/28709]\n",
            "loss: 1.261825  [ 5120/28709]\n",
            "loss: 1.235447  [ 5760/28709]\n",
            "loss: 1.177656  [ 6400/28709]\n",
            "loss: 1.152122  [ 7040/28709]\n",
            "loss: 1.184327  [ 7680/28709]\n",
            "loss: 1.166699  [ 8320/28709]\n",
            "loss: 1.226943  [ 8960/28709]\n",
            "loss: 1.204479  [ 9600/28709]\n",
            "loss: 1.076611  [10240/28709]\n",
            "loss: 1.183720  [10880/28709]\n",
            "loss: 1.142339  [11520/28709]\n",
            "loss: 1.203580  [12160/28709]\n",
            "loss: 1.211447  [12800/28709]\n",
            "loss: 1.244149  [13440/28709]\n",
            "loss: 1.164628  [14080/28709]\n",
            "loss: 1.274476  [14720/28709]\n",
            "loss: 1.141576  [15360/28709]\n",
            "loss: 1.058673  [16000/28709]\n",
            "loss: 1.231970  [16640/28709]\n",
            "loss: 1.274945  [17280/28709]\n",
            "loss: 1.164407  [17920/28709]\n",
            "loss: 1.182610  [18560/28709]\n",
            "loss: 1.143006  [19200/28709]\n",
            "loss: 0.970607  [19840/28709]\n",
            "loss: 1.211136  [20480/28709]\n",
            "loss: 1.132329  [21120/28709]\n",
            "loss: 1.316441  [21760/28709]\n",
            "loss: 1.321235  [22400/28709]\n",
            "loss: 1.146123  [23040/28709]\n",
            "loss: 1.257675  [23680/28709]\n",
            "loss: 1.283002  [24320/28709]\n",
            "loss: 1.317012  [24960/28709]\n",
            "loss: 1.212444  [25600/28709]\n",
            "loss: 1.304143  [26240/28709]\n",
            "loss: 1.222919  [26880/28709]\n",
            "loss: 1.271715  [27520/28709]\n",
            "loss: 1.141905  [28160/28709]\n",
            "\n",
            "Test Error\n",
            "\tAccuracy: 55.2%, Avg loss: 1.205215 \n",
            "\n",
            "Epoch 15\n",
            "-------------------------------\n",
            "loss: 1.342857  [    0/28709]\n",
            "loss: 1.153466  [  640/28709]\n",
            "loss: 1.260529  [ 1280/28709]\n",
            "loss: 1.346375  [ 1920/28709]\n",
            "loss: 1.137962  [ 2560/28709]\n",
            "loss: 1.262454  [ 3200/28709]\n",
            "loss: 1.212032  [ 3840/28709]\n",
            "loss: 0.952381  [ 4480/28709]\n",
            "loss: 0.980220  [ 5120/28709]\n",
            "loss: 1.341369  [ 5760/28709]\n",
            "loss: 1.085869  [ 6400/28709]\n",
            "loss: 1.117359  [ 7040/28709]\n",
            "loss: 1.197114  [ 7680/28709]\n",
            "loss: 1.165566  [ 8320/28709]\n",
            "loss: 1.315882  [ 8960/28709]\n",
            "loss: 1.130624  [ 9600/28709]\n",
            "loss: 1.138872  [10240/28709]\n",
            "loss: 1.122825  [10880/28709]\n",
            "loss: 1.161452  [11520/28709]\n",
            "loss: 0.946502  [12160/28709]\n",
            "loss: 1.293609  [12800/28709]\n",
            "loss: 1.155310  [13440/28709]\n",
            "loss: 1.149107  [14080/28709]\n",
            "loss: 1.185029  [14720/28709]\n",
            "loss: 1.114035  [15360/28709]\n",
            "loss: 1.068192  [16000/28709]\n",
            "loss: 1.336984  [16640/28709]\n",
            "loss: 1.133665  [17280/28709]\n",
            "loss: 1.160946  [17920/28709]\n",
            "loss: 1.164820  [18560/28709]\n",
            "loss: 1.299214  [19200/28709]\n",
            "loss: 1.015941  [19840/28709]\n",
            "loss: 1.137365  [20480/28709]\n",
            "loss: 1.326069  [21120/28709]\n",
            "loss: 1.081966  [21760/28709]\n",
            "loss: 1.276132  [22400/28709]\n",
            "loss: 1.101249  [23040/28709]\n",
            "loss: 1.077929  [23680/28709]\n",
            "loss: 1.377191  [24320/28709]\n",
            "loss: 1.030300  [24960/28709]\n",
            "loss: 1.284321  [25600/28709]\n",
            "loss: 1.403255  [26240/28709]\n",
            "loss: 1.587369  [26880/28709]\n",
            "loss: 1.051959  [27520/28709]\n",
            "loss: 1.311031  [28160/28709]\n",
            "\n",
            "Test Error\n",
            "\tAccuracy: 54.4%, Avg loss: 1.225172 \n",
            "\n",
            "Epoch 16\n",
            "-------------------------------\n",
            "loss: 1.159464  [    0/28709]\n",
            "loss: 1.073223  [  640/28709]\n",
            "loss: 1.202801  [ 1280/28709]\n",
            "loss: 1.251876  [ 1920/28709]\n",
            "loss: 1.136661  [ 2560/28709]\n",
            "loss: 1.218942  [ 3200/28709]\n",
            "loss: 1.076699  [ 3840/28709]\n",
            "loss: 1.185045  [ 4480/28709]\n",
            "loss: 1.058263  [ 5120/28709]\n",
            "loss: 1.521356  [ 5760/28709]\n",
            "loss: 1.182177  [ 6400/28709]\n",
            "loss: 1.113935  [ 7040/28709]\n",
            "loss: 1.208258  [ 7680/28709]\n",
            "loss: 1.242301  [ 8320/28709]\n",
            "loss: 0.964797  [ 8960/28709]\n",
            "loss: 1.164170  [ 9600/28709]\n",
            "loss: 0.943055  [10240/28709]\n",
            "loss: 1.072306  [10880/28709]\n",
            "loss: 1.116340  [11520/28709]\n",
            "loss: 0.944164  [12160/28709]\n",
            "loss: 1.241476  [12800/28709]\n",
            "loss: 1.158062  [13440/28709]\n",
            "loss: 1.075981  [14080/28709]\n",
            "loss: 1.066932  [14720/28709]\n",
            "loss: 1.262519  [15360/28709]\n",
            "loss: 1.285835  [16000/28709]\n",
            "loss: 1.227048  [16640/28709]\n",
            "loss: 1.088873  [17280/28709]\n",
            "loss: 1.137539  [17920/28709]\n",
            "loss: 1.246922  [18560/28709]\n",
            "loss: 1.076714  [19200/28709]\n",
            "loss: 1.066027  [19840/28709]\n",
            "loss: 1.107560  [20480/28709]\n",
            "loss: 1.210015  [21120/28709]\n",
            "loss: 1.074265  [21760/28709]\n",
            "loss: 1.185409  [22400/28709]\n",
            "loss: 0.994145  [23040/28709]\n",
            "loss: 1.253353  [23680/28709]\n",
            "loss: 1.067673  [24320/28709]\n",
            "loss: 1.075474  [24960/28709]\n",
            "loss: 1.013909  [25600/28709]\n",
            "loss: 1.243848  [26240/28709]\n",
            "loss: 0.932341  [26880/28709]\n",
            "loss: 1.249928  [27520/28709]\n",
            "loss: 1.101046  [28160/28709]\n",
            "\n",
            "Test Error\n",
            "\tAccuracy: 55.5%, Avg loss: 1.235200 \n",
            "\n",
            "Epoch 17\n",
            "-------------------------------\n",
            "loss: 0.937228  [    0/28709]\n",
            "loss: 1.038537  [  640/28709]\n",
            "loss: 1.241027  [ 1280/28709]\n",
            "loss: 1.207013  [ 1920/28709]\n",
            "loss: 1.256064  [ 2560/28709]\n",
            "loss: 1.050339  [ 3200/28709]\n",
            "loss: 1.035708  [ 3840/28709]\n",
            "loss: 1.159490  [ 4480/28709]\n",
            "loss: 0.974423  [ 5120/28709]\n",
            "loss: 1.235046  [ 5760/28709]\n",
            "loss: 1.674462  [ 6400/28709]\n",
            "loss: 1.238571  [ 7040/28709]\n",
            "loss: 1.061703  [ 7680/28709]\n",
            "loss: 1.274446  [ 8320/28709]\n",
            "loss: 1.064402  [ 8960/28709]\n",
            "loss: 1.143703  [ 9600/28709]\n",
            "loss: 1.171801  [10240/28709]\n",
            "loss: 1.064729  [10880/28709]\n",
            "loss: 1.046015  [11520/28709]\n",
            "loss: 1.150416  [12160/28709]\n",
            "loss: 1.005601  [12800/28709]\n",
            "loss: 1.083710  [13440/28709]\n",
            "loss: 0.962030  [14080/28709]\n",
            "loss: 1.011578  [14720/28709]\n",
            "loss: 1.153261  [15360/28709]\n",
            "loss: 1.038270  [16000/28709]\n",
            "loss: 1.169216  [16640/28709]\n",
            "loss: 1.231515  [17280/28709]\n",
            "loss: 1.084693  [17920/28709]\n",
            "loss: 1.027129  [18560/28709]\n",
            "loss: 1.061852  [19200/28709]\n",
            "loss: 0.970322  [19840/28709]\n",
            "loss: 1.215544  [20480/28709]\n",
            "loss: 1.137496  [21120/28709]\n",
            "loss: 1.042745  [21760/28709]\n",
            "loss: 1.133667  [22400/28709]\n",
            "loss: 1.202333  [23040/28709]\n",
            "loss: 1.129405  [23680/28709]\n",
            "loss: 1.200846  [24320/28709]\n",
            "loss: 1.321817  [24960/28709]\n",
            "loss: 1.122812  [25600/28709]\n",
            "loss: 1.078622  [26240/28709]\n",
            "loss: 1.325884  [26880/28709]\n",
            "loss: 1.432164  [27520/28709]\n",
            "loss: 1.131663  [28160/28709]\n",
            "\n",
            "Test Error\n",
            "\tAccuracy: 53.2%, Avg loss: 1.261220 \n",
            "\n",
            "Epoch 18\n",
            "-------------------------------\n",
            "loss: 1.075606  [    0/28709]\n",
            "loss: 1.087592  [  640/28709]\n",
            "loss: 1.176562  [ 1280/28709]\n",
            "loss: 1.276495  [ 1920/28709]\n",
            "loss: 1.102141  [ 2560/28709]\n",
            "loss: 1.171443  [ 3200/28709]\n",
            "loss: 1.178967  [ 3840/28709]\n",
            "loss: 1.086312  [ 4480/28709]\n",
            "loss: 1.296733  [ 5120/28709]\n",
            "loss: 1.124250  [ 5760/28709]\n",
            "loss: 1.336010  [ 6400/28709]\n",
            "loss: 1.182516  [ 7040/28709]\n",
            "loss: 1.382722  [ 7680/28709]\n",
            "loss: 1.150693  [ 8320/28709]\n",
            "loss: 1.053390  [ 8960/28709]\n",
            "loss: 1.128402  [ 9600/28709]\n",
            "loss: 1.239614  [10240/28709]\n",
            "loss: 1.095272  [10880/28709]\n",
            "loss: 0.964528  [11520/28709]\n",
            "loss: 1.149036  [12160/28709]\n",
            "loss: 1.146606  [12800/28709]\n",
            "loss: 1.048305  [13440/28709]\n",
            "loss: 1.253816  [14080/28709]\n",
            "loss: 1.116693  [14720/28709]\n",
            "loss: 0.951975  [15360/28709]\n",
            "loss: 1.140969  [16000/28709]\n",
            "loss: 1.245881  [16640/28709]\n",
            "loss: 1.036487  [17280/28709]\n",
            "loss: 1.147628  [17920/28709]\n",
            "loss: 1.155925  [18560/28709]\n",
            "loss: 1.091539  [19200/28709]\n",
            "loss: 1.243142  [19840/28709]\n",
            "loss: 1.254184  [20480/28709]\n",
            "loss: 1.234942  [21120/28709]\n",
            "loss: 1.114024  [21760/28709]\n",
            "loss: 1.123777  [22400/28709]\n",
            "loss: 1.137204  [23040/28709]\n",
            "loss: 1.264731  [23680/28709]\n",
            "loss: 1.037687  [24320/28709]\n",
            "loss: 1.294811  [24960/28709]\n",
            "loss: 1.043398  [25600/28709]\n",
            "loss: 1.192149  [26240/28709]\n",
            "loss: 1.342597  [26880/28709]\n",
            "loss: 0.968474  [27520/28709]\n",
            "loss: 1.244044  [28160/28709]\n",
            "\n",
            "Test Error\n",
            "\tAccuracy: 56.2%, Avg loss: 1.187210 \n",
            "\n",
            "Epoch 19\n",
            "-------------------------------\n",
            "loss: 1.064193  [    0/28709]\n",
            "loss: 0.928008  [  640/28709]\n",
            "loss: 1.244837  [ 1280/28709]\n",
            "loss: 0.953864  [ 1920/28709]\n",
            "loss: 0.999993  [ 2560/28709]\n",
            "loss: 1.094733  [ 3200/28709]\n",
            "loss: 0.983944  [ 3840/28709]\n",
            "loss: 1.170950  [ 4480/28709]\n",
            "loss: 0.961456  [ 5120/28709]\n",
            "loss: 1.148132  [ 5760/28709]\n",
            "loss: 1.073911  [ 6400/28709]\n",
            "loss: 0.954675  [ 7040/28709]\n",
            "loss: 1.113826  [ 7680/28709]\n",
            "loss: 1.144535  [ 8320/28709]\n",
            "loss: 1.199507  [ 8960/28709]\n",
            "loss: 1.145112  [ 9600/28709]\n",
            "loss: 1.083210  [10240/28709]\n",
            "loss: 0.951084  [10880/28709]\n",
            "loss: 0.814191  [11520/28709]\n",
            "loss: 0.975981  [12160/28709]\n",
            "loss: 1.219115  [12800/28709]\n",
            "loss: 1.317948  [13440/28709]\n",
            "loss: 1.198449  [14080/28709]\n",
            "loss: 1.085480  [14720/28709]\n",
            "loss: 0.972260  [15360/28709]\n",
            "loss: 1.207265  [16000/28709]\n",
            "loss: 0.869297  [16640/28709]\n",
            "loss: 0.942000  [17280/28709]\n",
            "loss: 1.139260  [17920/28709]\n",
            "loss: 1.245947  [18560/28709]\n",
            "loss: 1.146438  [19200/28709]\n",
            "loss: 1.048784  [19840/28709]\n",
            "loss: 1.028703  [20480/28709]\n",
            "loss: 1.088770  [21120/28709]\n",
            "loss: 0.993155  [21760/28709]\n",
            "loss: 0.910438  [22400/28709]\n",
            "loss: 1.103461  [23040/28709]\n",
            "loss: 0.983101  [23680/28709]\n",
            "loss: 1.090832  [24320/28709]\n",
            "loss: 0.864418  [24960/28709]\n",
            "loss: 1.035262  [25600/28709]\n",
            "loss: 1.093732  [26240/28709]\n",
            "loss: 1.023429  [26880/28709]\n",
            "loss: 0.908024  [27520/28709]\n",
            "loss: 1.208058  [28160/28709]\n",
            "\n",
            "Test Error\n",
            "\tAccuracy: 56.8%, Avg loss: 1.181690 \n",
            "\n",
            "Epoch 20\n",
            "-------------------------------\n",
            "loss: 0.989143  [    0/28709]\n",
            "loss: 0.987373  [  640/28709]\n",
            "loss: 1.009185  [ 1280/28709]\n",
            "loss: 0.918736  [ 1920/28709]\n",
            "loss: 0.980198  [ 2560/28709]\n",
            "loss: 1.300722  [ 3200/28709]\n",
            "loss: 0.984107  [ 3840/28709]\n",
            "loss: 0.861246  [ 4480/28709]\n",
            "loss: 1.314718  [ 5120/28709]\n",
            "loss: 1.036027  [ 5760/28709]\n",
            "loss: 1.182231  [ 6400/28709]\n",
            "loss: 0.958922  [ 7040/28709]\n",
            "loss: 1.052932  [ 7680/28709]\n",
            "loss: 1.013389  [ 8320/28709]\n",
            "loss: 1.041862  [ 8960/28709]\n",
            "loss: 1.105925  [ 9600/28709]\n",
            "loss: 1.036808  [10240/28709]\n",
            "loss: 1.142362  [10880/28709]\n",
            "loss: 0.945933  [11520/28709]\n",
            "loss: 0.947086  [12160/28709]\n",
            "loss: 0.937674  [12800/28709]\n",
            "loss: 1.005530  [13440/28709]\n",
            "loss: 1.076387  [14080/28709]\n",
            "loss: 1.151904  [14720/28709]\n",
            "loss: 1.066651  [15360/28709]\n",
            "loss: 1.036802  [16000/28709]\n",
            "loss: 1.036792  [16640/28709]\n",
            "loss: 1.185282  [17280/28709]\n",
            "loss: 1.040242  [17920/28709]\n",
            "loss: 0.919130  [18560/28709]\n",
            "loss: 1.030074  [19200/28709]\n",
            "loss: 1.150411  [19840/28709]\n",
            "loss: 1.064733  [20480/28709]\n",
            "loss: 0.825807  [21120/28709]\n",
            "loss: 1.236063  [21760/28709]\n",
            "loss: 0.999515  [22400/28709]\n",
            "loss: 1.281053  [23040/28709]\n",
            "loss: 1.012491  [23680/28709]\n",
            "loss: 1.110499  [24320/28709]\n",
            "loss: 1.032561  [24960/28709]\n",
            "loss: 1.172362  [25600/28709]\n",
            "loss: 1.124942  [26240/28709]\n",
            "loss: 1.172602  [26880/28709]\n",
            "loss: 1.107467  [27520/28709]\n",
            "loss: 1.089703  [28160/28709]\n",
            "\n",
            "Test Error\n",
            "\tAccuracy: 56.5%, Avg loss: 1.181124 \n",
            "\n",
            "Epoch 21\n",
            "-------------------------------\n",
            "loss: 0.962475  [    0/28709]\n",
            "loss: 0.918068  [  640/28709]\n",
            "loss: 0.648526  [ 1280/28709]\n",
            "loss: 0.980153  [ 1920/28709]\n",
            "loss: 1.168191  [ 2560/28709]\n",
            "loss: 1.117059  [ 3200/28709]\n",
            "loss: 1.011865  [ 3840/28709]\n",
            "loss: 1.053283  [ 4480/28709]\n",
            "loss: 0.970213  [ 5120/28709]\n",
            "loss: 1.022466  [ 5760/28709]\n",
            "loss: 1.381129  [ 6400/28709]\n",
            "loss: 0.951809  [ 7040/28709]\n",
            "loss: 1.091725  [ 7680/28709]\n",
            "loss: 1.048435  [ 8320/28709]\n",
            "loss: 1.007953  [ 8960/28709]\n",
            "loss: 1.105678  [ 9600/28709]\n",
            "loss: 0.970351  [10240/28709]\n",
            "loss: 0.899960  [10880/28709]\n",
            "loss: 1.103063  [11520/28709]\n",
            "loss: 1.208550  [12160/28709]\n",
            "loss: 1.183282  [12800/28709]\n",
            "loss: 1.201811  [13440/28709]\n",
            "loss: 0.927986  [14080/28709]\n",
            "loss: 0.951883  [14720/28709]\n",
            "loss: 1.024237  [15360/28709]\n",
            "loss: 0.998660  [16000/28709]\n",
            "loss: 0.937845  [16640/28709]\n",
            "loss: 1.049299  [17280/28709]\n",
            "loss: 1.001145  [17920/28709]\n",
            "loss: 1.047887  [18560/28709]\n",
            "loss: 1.119878  [19200/28709]\n",
            "loss: 0.890680  [19840/28709]\n",
            "loss: 1.162545  [20480/28709]\n",
            "loss: 1.062685  [21120/28709]\n",
            "loss: 1.079382  [21760/28709]\n",
            "loss: 0.909340  [22400/28709]\n",
            "loss: 1.076888  [23040/28709]\n",
            "loss: 0.956512  [23680/28709]\n",
            "loss: 0.934971  [24320/28709]\n",
            "loss: 1.199315  [24960/28709]\n",
            "loss: 1.115168  [25600/28709]\n",
            "loss: 0.917563  [26240/28709]\n",
            "loss: 1.003063  [26880/28709]\n",
            "loss: 1.092551  [27520/28709]\n",
            "loss: 1.075957  [28160/28709]\n",
            "\n",
            "Test Error\n",
            "\tAccuracy: 57.5%, Avg loss: 1.167577 \n",
            "\n",
            "Epoch 22\n",
            "-------------------------------\n",
            "loss: 0.934990  [    0/28709]\n",
            "loss: 1.049191  [  640/28709]\n",
            "loss: 0.945252  [ 1280/28709]\n",
            "loss: 1.016718  [ 1920/28709]\n",
            "loss: 0.968917  [ 2560/28709]\n",
            "loss: 0.913510  [ 3200/28709]\n",
            "loss: 0.991087  [ 3840/28709]\n",
            "loss: 1.085104  [ 4480/28709]\n",
            "loss: 0.894452  [ 5120/28709]\n",
            "loss: 0.905168  [ 5760/28709]\n",
            "loss: 1.261153  [ 6400/28709]\n",
            "loss: 0.794097  [ 7040/28709]\n",
            "loss: 0.953410  [ 7680/28709]\n",
            "loss: 0.934129  [ 8320/28709]\n",
            "loss: 0.858983  [ 8960/28709]\n",
            "loss: 1.270101  [ 9600/28709]\n",
            "loss: 0.949709  [10240/28709]\n",
            "loss: 1.075437  [10880/28709]\n",
            "loss: 0.953408  [11520/28709]\n",
            "loss: 0.961642  [12160/28709]\n",
            "loss: 0.802027  [12800/28709]\n",
            "loss: 0.801397  [13440/28709]\n",
            "loss: 0.953370  [14080/28709]\n",
            "loss: 0.883090  [14720/28709]\n",
            "loss: 0.890763  [15360/28709]\n",
            "loss: 1.106697  [16000/28709]\n",
            "loss: 1.129387  [16640/28709]\n",
            "loss: 1.067547  [17280/28709]\n",
            "loss: 1.115860  [17920/28709]\n",
            "loss: 1.274897  [18560/28709]\n",
            "loss: 1.084495  [19200/28709]\n",
            "loss: 1.065616  [19840/28709]\n",
            "loss: 0.995138  [20480/28709]\n",
            "loss: 0.975933  [21120/28709]\n",
            "loss: 1.013102  [21760/28709]\n",
            "loss: 0.992009  [22400/28709]\n",
            "loss: 0.974702  [23040/28709]\n",
            "loss: 0.956065  [23680/28709]\n",
            "loss: 0.998439  [24320/28709]\n",
            "loss: 1.060716  [24960/28709]\n",
            "loss: 0.998174  [25600/28709]\n",
            "loss: 1.111472  [26240/28709]\n",
            "loss: 1.034114  [26880/28709]\n",
            "loss: 1.289903  [27520/28709]\n",
            "loss: 0.861989  [28160/28709]\n",
            "\n",
            "Test Error\n",
            "\tAccuracy: 57.5%, Avg loss: 1.216058 \n",
            "\n",
            "Epoch 23\n",
            "-------------------------------\n",
            "loss: 0.879112  [    0/28709]\n",
            "loss: 0.979463  [  640/28709]\n",
            "loss: 0.888850  [ 1280/28709]\n",
            "loss: 0.870298  [ 1920/28709]\n",
            "loss: 0.997689  [ 2560/28709]\n",
            "loss: 1.115068  [ 3200/28709]\n",
            "loss: 0.980551  [ 3840/28709]\n",
            "loss: 1.150279  [ 4480/28709]\n",
            "loss: 1.050515  [ 5120/28709]\n",
            "loss: 0.914492  [ 5760/28709]\n",
            "loss: 0.895562  [ 6400/28709]\n",
            "loss: 0.952167  [ 7040/28709]\n",
            "loss: 0.951025  [ 7680/28709]\n",
            "loss: 0.900578  [ 8320/28709]\n",
            "loss: 0.895934  [ 8960/28709]\n",
            "loss: 1.059016  [ 9600/28709]\n",
            "loss: 0.998417  [10240/28709]\n",
            "loss: 1.111730  [10880/28709]\n",
            "loss: 0.937994  [11520/28709]\n",
            "loss: 0.830408  [12160/28709]\n",
            "loss: 0.950458  [12800/28709]\n",
            "loss: 1.046725  [13440/28709]\n",
            "loss: 1.083204  [14080/28709]\n",
            "loss: 1.113229  [14720/28709]\n",
            "loss: 1.343462  [15360/28709]\n",
            "loss: 1.194256  [16000/28709]\n",
            "loss: 0.845592  [16640/28709]\n",
            "loss: 1.100603  [17280/28709]\n",
            "loss: 0.922693  [17920/28709]\n",
            "loss: 1.143013  [18560/28709]\n",
            "loss: 0.777465  [19200/28709]\n",
            "loss: 0.849892  [19840/28709]\n",
            "loss: 0.964056  [20480/28709]\n",
            "loss: 1.161962  [21120/28709]\n",
            "loss: 1.103836  [21760/28709]\n",
            "loss: 0.880834  [22400/28709]\n",
            "loss: 0.958459  [23040/28709]\n",
            "loss: 0.873125  [23680/28709]\n",
            "loss: 0.856335  [24320/28709]\n",
            "loss: 1.170678  [24960/28709]\n",
            "loss: 1.135981  [25600/28709]\n",
            "loss: 1.036033  [26240/28709]\n",
            "loss: 1.281233  [26880/28709]\n",
            "loss: 1.108213  [27520/28709]\n",
            "loss: 1.007932  [28160/28709]\n",
            "\n",
            "Test Error\n",
            "\tAccuracy: 57.3%, Avg loss: 1.171014 \n",
            "\n",
            "Epoch 24\n",
            "-------------------------------\n",
            "loss: 0.933636  [    0/28709]\n",
            "loss: 1.032241  [  640/28709]\n",
            "loss: 0.926074  [ 1280/28709]\n",
            "loss: 1.070173  [ 1920/28709]\n",
            "loss: 0.793941  [ 2560/28709]\n",
            "loss: 0.854133  [ 3200/28709]\n",
            "loss: 1.096985  [ 3840/28709]\n",
            "loss: 0.888535  [ 4480/28709]\n",
            "loss: 0.895174  [ 5120/28709]\n",
            "loss: 0.848493  [ 5760/28709]\n",
            "loss: 0.923351  [ 6400/28709]\n",
            "loss: 1.002504  [ 7040/28709]\n",
            "loss: 1.034323  [ 7680/28709]\n",
            "loss: 1.092075  [ 8320/28709]\n",
            "loss: 1.023190  [ 8960/28709]\n",
            "loss: 0.649951  [ 9600/28709]\n",
            "loss: 1.010990  [10240/28709]\n",
            "loss: 1.013941  [10880/28709]\n",
            "loss: 0.932321  [11520/28709]\n",
            "loss: 0.780336  [12160/28709]\n",
            "loss: 1.028508  [12800/28709]\n",
            "loss: 1.047735  [13440/28709]\n",
            "loss: 1.045928  [14080/28709]\n",
            "loss: 0.900491  [14720/28709]\n",
            "loss: 0.921440  [15360/28709]\n",
            "loss: 0.977194  [16000/28709]\n",
            "loss: 0.945603  [16640/28709]\n",
            "loss: 0.990568  [17280/28709]\n",
            "loss: 0.954456  [17920/28709]\n",
            "loss: 1.102697  [18560/28709]\n",
            "loss: 0.872536  [19200/28709]\n",
            "loss: 1.005732  [19840/28709]\n",
            "loss: 0.977886  [20480/28709]\n",
            "loss: 0.895882  [21120/28709]\n",
            "loss: 1.056736  [21760/28709]\n",
            "loss: 1.130348  [22400/28709]\n",
            "loss: 0.797317  [23040/28709]\n",
            "loss: 0.762819  [23680/28709]\n",
            "loss: 0.892236  [24320/28709]\n",
            "loss: 0.944020  [24960/28709]\n",
            "loss: 0.916706  [25600/28709]\n",
            "loss: 0.864740  [26240/28709]\n",
            "loss: 0.931349  [26880/28709]\n",
            "loss: 1.106991  [27520/28709]\n",
            "loss: 1.126545  [28160/28709]\n",
            "\n",
            "Test Error\n",
            "\tAccuracy: 56.9%, Avg loss: 1.167904 \n",
            "\n",
            "Epoch 25\n",
            "-------------------------------\n",
            "loss: 0.950000  [    0/28709]\n",
            "loss: 0.998949  [  640/28709]\n",
            "loss: 0.832688  [ 1280/28709]\n",
            "loss: 0.910347  [ 1920/28709]\n",
            "loss: 1.013982  [ 2560/28709]\n",
            "loss: 0.910565  [ 3200/28709]\n",
            "loss: 0.776359  [ 3840/28709]\n",
            "loss: 0.737564  [ 4480/28709]\n",
            "loss: 0.783794  [ 5120/28709]\n",
            "loss: 0.758447  [ 5760/28709]\n",
            "loss: 0.727149  [ 6400/28709]\n",
            "loss: 0.909366  [ 7040/28709]\n",
            "loss: 1.148265  [ 7680/28709]\n",
            "loss: 0.953939  [ 8320/28709]\n",
            "loss: 0.825248  [ 8960/28709]\n",
            "loss: 0.778106  [ 9600/28709]\n",
            "loss: 1.048385  [10240/28709]\n",
            "loss: 0.865613  [10880/28709]\n",
            "loss: 0.991559  [11520/28709]\n",
            "loss: 0.972028  [12160/28709]\n",
            "loss: 1.026128  [12800/28709]\n",
            "loss: 0.911722  [13440/28709]\n",
            "loss: 1.016050  [14080/28709]\n",
            "loss: 0.762875  [14720/28709]\n",
            "loss: 1.153715  [15360/28709]\n",
            "loss: 0.855781  [16000/28709]\n",
            "loss: 0.973835  [16640/28709]\n",
            "loss: 0.845537  [17280/28709]\n",
            "loss: 0.900955  [17920/28709]\n",
            "loss: 0.830150  [18560/28709]\n",
            "loss: 0.866577  [19200/28709]\n",
            "loss: 0.998991  [19840/28709]\n",
            "loss: 0.936436  [20480/28709]\n",
            "loss: 1.026597  [21120/28709]\n",
            "loss: 0.865818  [21760/28709]\n",
            "loss: 1.099409  [22400/28709]\n",
            "loss: 0.894192  [23040/28709]\n",
            "loss: 1.064139  [23680/28709]\n",
            "loss: 1.020778  [24320/28709]\n",
            "loss: 0.810427  [24960/28709]\n",
            "loss: 0.852775  [25600/28709]\n",
            "loss: 0.887758  [26240/28709]\n",
            "loss: 0.798656  [26880/28709]\n",
            "loss: 0.869956  [27520/28709]\n",
            "loss: 1.141332  [28160/28709]\n",
            "\n",
            "Test Error\n",
            "\tAccuracy: 57.9%, Avg loss: 1.181676 \n",
            "\n",
            "Epoch 26\n",
            "-------------------------------\n",
            "loss: 0.815725  [    0/28709]\n",
            "loss: 0.952578  [  640/28709]\n",
            "loss: 0.943680  [ 1280/28709]\n",
            "loss: 1.142747  [ 1920/28709]\n",
            "loss: 0.882190  [ 2560/28709]\n",
            "loss: 0.969091  [ 3200/28709]\n",
            "loss: 0.980578  [ 3840/28709]\n",
            "loss: 1.152007  [ 4480/28709]\n",
            "loss: 0.860132  [ 5120/28709]\n",
            "loss: 0.857665  [ 5760/28709]\n",
            "loss: 1.022883  [ 6400/28709]\n",
            "loss: 0.952114  [ 7040/28709]\n",
            "loss: 0.956106  [ 7680/28709]\n",
            "loss: 0.765847  [ 8320/28709]\n",
            "loss: 1.037250  [ 8960/28709]\n",
            "loss: 1.089716  [ 9600/28709]\n",
            "loss: 0.962264  [10240/28709]\n",
            "loss: 0.785378  [10880/28709]\n",
            "loss: 0.870945  [11520/28709]\n",
            "loss: 1.028604  [12160/28709]\n",
            "loss: 0.917620  [12800/28709]\n",
            "loss: 0.837480  [13440/28709]\n",
            "loss: 0.706944  [14080/28709]\n",
            "loss: 0.788134  [14720/28709]\n",
            "loss: 1.041067  [15360/28709]\n",
            "loss: 1.010359  [16000/28709]\n",
            "loss: 0.941648  [16640/28709]\n",
            "loss: 0.784709  [17280/28709]\n",
            "loss: 1.102696  [17920/28709]\n",
            "loss: 0.991589  [18560/28709]\n",
            "loss: 0.807183  [19200/28709]\n",
            "loss: 0.896095  [19840/28709]\n",
            "loss: 1.018089  [20480/28709]\n",
            "loss: 0.878927  [21120/28709]\n",
            "loss: 0.955315  [21760/28709]\n",
            "loss: 0.940195  [22400/28709]\n",
            "loss: 1.149768  [23040/28709]\n",
            "loss: 1.141197  [23680/28709]\n",
            "loss: 0.801505  [24320/28709]\n",
            "loss: 0.867929  [24960/28709]\n",
            "loss: 0.898710  [25600/28709]\n",
            "loss: 0.979554  [26240/28709]\n",
            "loss: 0.930096  [26880/28709]\n",
            "loss: 1.061785  [27520/28709]\n",
            "loss: 0.898332  [28160/28709]\n",
            "\n",
            "Test Error\n",
            "\tAccuracy: 57.2%, Avg loss: 1.229185 \n",
            "\n",
            "Epoch 27\n",
            "-------------------------------\n",
            "loss: 0.773794  [    0/28709]\n",
            "loss: 0.880070  [  640/28709]\n",
            "loss: 1.061832  [ 1280/28709]\n",
            "loss: 0.822191  [ 1920/28709]\n",
            "loss: 1.082416  [ 2560/28709]\n",
            "loss: 0.939263  [ 3200/28709]\n",
            "loss: 0.831366  [ 3840/28709]\n",
            "loss: 0.952848  [ 4480/28709]\n",
            "loss: 0.976128  [ 5120/28709]\n",
            "loss: 1.029220  [ 5760/28709]\n",
            "loss: 1.037408  [ 6400/28709]\n",
            "loss: 0.946313  [ 7040/28709]\n",
            "loss: 0.942138  [ 7680/28709]\n",
            "loss: 0.629898  [ 8320/28709]\n",
            "loss: 0.989873  [ 8960/28709]\n",
            "loss: 0.825587  [ 9600/28709]\n",
            "loss: 0.948313  [10240/28709]\n",
            "loss: 0.969536  [10880/28709]\n",
            "loss: 0.994798  [11520/28709]\n",
            "loss: 0.902862  [12160/28709]\n",
            "loss: 0.827429  [12800/28709]\n",
            "loss: 0.945566  [13440/28709]\n",
            "loss: 0.791445  [14080/28709]\n",
            "loss: 1.056372  [14720/28709]\n",
            "loss: 0.929548  [15360/28709]\n",
            "loss: 0.873473  [16000/28709]\n",
            "loss: 1.044786  [16640/28709]\n",
            "loss: 1.129714  [17280/28709]\n",
            "loss: 0.880105  [17920/28709]\n",
            "loss: 0.765843  [18560/28709]\n",
            "loss: 1.039769  [19200/28709]\n",
            "loss: 0.765111  [19840/28709]\n",
            "loss: 0.816487  [20480/28709]\n",
            "loss: 0.962291  [21120/28709]\n",
            "loss: 0.845168  [21760/28709]\n",
            "loss: 1.096764  [22400/28709]\n",
            "loss: 0.746342  [23040/28709]\n",
            "loss: 0.894351  [23680/28709]\n",
            "loss: 1.122094  [24320/28709]\n",
            "loss: 0.720591  [24960/28709]\n",
            "loss: 0.816958  [25600/28709]\n",
            "loss: 0.726782  [26240/28709]\n",
            "loss: 1.104023  [26880/28709]\n",
            "loss: 1.061177  [27520/28709]\n",
            "loss: 0.869110  [28160/28709]\n",
            "\n",
            "Test Error\n",
            "\tAccuracy: 58.3%, Avg loss: 1.160953 \n",
            "\n",
            "Epoch 28\n",
            "-------------------------------\n",
            "loss: 0.914414  [    0/28709]\n",
            "loss: 0.996628  [  640/28709]\n",
            "loss: 0.887970  [ 1280/28709]\n",
            "loss: 0.997494  [ 1920/28709]\n",
            "loss: 0.957036  [ 2560/28709]\n",
            "loss: 0.780566  [ 3200/28709]\n",
            "loss: 0.806019  [ 3840/28709]\n",
            "loss: 0.986710  [ 4480/28709]\n",
            "loss: 0.886818  [ 5120/28709]\n",
            "loss: 0.814653  [ 5760/28709]\n",
            "loss: 0.818587  [ 6400/28709]\n",
            "loss: 0.735471  [ 7040/28709]\n",
            "loss: 1.210157  [ 7680/28709]\n",
            "loss: 0.930981  [ 8320/28709]\n",
            "loss: 0.869691  [ 8960/28709]\n",
            "loss: 0.802504  [ 9600/28709]\n",
            "loss: 1.090789  [10240/28709]\n",
            "loss: 0.935809  [10880/28709]\n",
            "loss: 0.871397  [11520/28709]\n",
            "loss: 0.831719  [12160/28709]\n",
            "loss: 0.794397  [12800/28709]\n",
            "loss: 0.877101  [13440/28709]\n",
            "loss: 0.773722  [14080/28709]\n",
            "loss: 0.779187  [14720/28709]\n",
            "loss: 0.805963  [15360/28709]\n",
            "loss: 0.942140  [16000/28709]\n",
            "loss: 0.984640  [16640/28709]\n",
            "loss: 0.801454  [17280/28709]\n",
            "loss: 0.827399  [17920/28709]\n",
            "loss: 0.569795  [18560/28709]\n",
            "loss: 0.872964  [19200/28709]\n",
            "loss: 0.788176  [19840/28709]\n",
            "loss: 0.974512  [20480/28709]\n",
            "loss: 0.939903  [21120/28709]\n",
            "loss: 0.762889  [21760/28709]\n",
            "loss: 1.021198  [22400/28709]\n",
            "loss: 0.660406  [23040/28709]\n",
            "loss: 1.076295  [23680/28709]\n",
            "loss: 0.972691  [24320/28709]\n",
            "loss: 0.760764  [24960/28709]\n",
            "loss: 1.056815  [25600/28709]\n",
            "loss: 0.874617  [26240/28709]\n",
            "loss: 0.782863  [26880/28709]\n",
            "loss: 1.162558  [27520/28709]\n",
            "loss: 1.111993  [28160/28709]\n",
            "\n",
            "Test Error\n",
            "\tAccuracy: 58.0%, Avg loss: 1.198047 \n",
            "\n",
            "Epoch 29\n",
            "-------------------------------\n",
            "loss: 0.984309  [    0/28709]\n",
            "loss: 0.813739  [  640/28709]\n",
            "loss: 1.106980  [ 1280/28709]\n",
            "loss: 0.935442  [ 1920/28709]\n",
            "loss: 0.828549  [ 2560/28709]\n",
            "loss: 0.704274  [ 3200/28709]\n",
            "loss: 0.869556  [ 3840/28709]\n",
            "loss: 0.968150  [ 4480/28709]\n",
            "loss: 0.930282  [ 5120/28709]\n",
            "loss: 0.876814  [ 5760/28709]\n",
            "loss: 0.845297  [ 6400/28709]\n",
            "loss: 0.902425  [ 7040/28709]\n",
            "loss: 0.815292  [ 7680/28709]\n",
            "loss: 0.927335  [ 8320/28709]\n",
            "loss: 0.820031  [ 8960/28709]\n",
            "loss: 1.131242  [ 9600/28709]\n",
            "loss: 1.062132  [10240/28709]\n",
            "loss: 0.873027  [10880/28709]\n",
            "loss: 1.068367  [11520/28709]\n",
            "loss: 1.056303  [12160/28709]\n",
            "loss: 0.885362  [12800/28709]\n",
            "loss: 0.908289  [13440/28709]\n",
            "loss: 0.916436  [14080/28709]\n",
            "loss: 0.962819  [14720/28709]\n",
            "loss: 1.032956  [15360/28709]\n",
            "loss: 0.927962  [16000/28709]\n",
            "loss: 0.807754  [16640/28709]\n",
            "loss: 0.870274  [17280/28709]\n",
            "loss: 0.760615  [17920/28709]\n",
            "loss: 0.720800  [18560/28709]\n",
            "loss: 0.866548  [19200/28709]\n",
            "loss: 1.041416  [19840/28709]\n",
            "loss: 1.035267  [20480/28709]\n",
            "loss: 0.982255  [21120/28709]\n",
            "loss: 0.896388  [21760/28709]\n",
            "loss: 0.995461  [22400/28709]\n",
            "loss: 0.726127  [23040/28709]\n",
            "loss: 1.017331  [23680/28709]\n",
            "loss: 0.952115  [24320/28709]\n",
            "loss: 1.044890  [24960/28709]\n",
            "loss: 0.885516  [25600/28709]\n",
            "loss: 0.890830  [26240/28709]\n",
            "loss: 0.902915  [26880/28709]\n",
            "loss: 0.725839  [27520/28709]\n",
            "loss: 0.984890  [28160/28709]\n",
            "\n",
            "Test Error\n",
            "\tAccuracy: 58.1%, Avg loss: 1.197233 \n",
            "\n",
            "Epoch 30\n",
            "-------------------------------\n",
            "loss: 0.754522  [    0/28709]\n",
            "loss: 0.778476  [  640/28709]\n",
            "loss: 0.785021  [ 1280/28709]\n",
            "loss: 0.683310  [ 1920/28709]\n",
            "loss: 0.967547  [ 2560/28709]\n",
            "loss: 0.897509  [ 3200/28709]\n",
            "loss: 0.922191  [ 3840/28709]\n",
            "loss: 0.935857  [ 4480/28709]\n",
            "loss: 0.912907  [ 5120/28709]\n",
            "loss: 0.928251  [ 5760/28709]\n",
            "loss: 0.919473  [ 6400/28709]\n",
            "loss: 0.885737  [ 7040/28709]\n",
            "loss: 0.691273  [ 7680/28709]\n",
            "loss: 0.796728  [ 8320/28709]\n",
            "loss: 1.017483  [ 8960/28709]\n",
            "loss: 0.852546  [ 9600/28709]\n",
            "loss: 0.854394  [10240/28709]\n",
            "loss: 0.941808  [10880/28709]\n",
            "loss: 0.997719  [11520/28709]\n",
            "loss: 0.880959  [12160/28709]\n",
            "loss: 0.963620  [12800/28709]\n",
            "loss: 0.851917  [13440/28709]\n",
            "loss: 0.905662  [14080/28709]\n",
            "loss: 0.911658  [14720/28709]\n",
            "loss: 1.087721  [15360/28709]\n",
            "loss: 0.883037  [16000/28709]\n",
            "loss: 1.019550  [16640/28709]\n",
            "loss: 0.656774  [17280/28709]\n",
            "loss: 1.011851  [17920/28709]\n",
            "loss: 0.786532  [18560/28709]\n",
            "loss: 0.878710  [19200/28709]\n",
            "loss: 1.203318  [19840/28709]\n",
            "loss: 0.860404  [20480/28709]\n",
            "loss: 0.993809  [21120/28709]\n",
            "loss: 1.045342  [21760/28709]\n",
            "loss: 0.772177  [22400/28709]\n",
            "loss: 1.012466  [23040/28709]\n",
            "loss: 0.919064  [23680/28709]\n",
            "loss: 0.826430  [24320/28709]\n",
            "loss: 0.970342  [24960/28709]\n",
            "loss: 0.831171  [25600/28709]\n",
            "loss: 1.019684  [26240/28709]\n",
            "loss: 0.670093  [26880/28709]\n",
            "loss: 0.811717  [27520/28709]\n",
            "loss: 0.848860  [28160/28709]\n",
            "\n",
            "Test Error\n",
            "\tAccuracy: 57.8%, Avg loss: 1.179579 \n",
            "\n",
            "Epoch 31\n",
            "-------------------------------\n",
            "loss: 1.203616  [    0/28709]\n",
            "loss: 0.751962  [  640/28709]\n",
            "loss: 1.102038  [ 1280/28709]\n",
            "loss: 1.097027  [ 1920/28709]\n",
            "loss: 0.938389  [ 2560/28709]\n",
            "loss: 0.679163  [ 3200/28709]\n",
            "loss: 0.952384  [ 3840/28709]\n",
            "loss: 0.769255  [ 4480/28709]\n",
            "loss: 1.060657  [ 5120/28709]\n",
            "loss: 0.963794  [ 5760/28709]\n",
            "loss: 0.936143  [ 6400/28709]\n",
            "loss: 0.781607  [ 7040/28709]\n",
            "loss: 0.833718  [ 7680/28709]\n",
            "loss: 0.720331  [ 8320/28709]\n",
            "loss: 0.705914  [ 8960/28709]\n",
            "loss: 1.008709  [ 9600/28709]\n",
            "loss: 1.031523  [10240/28709]\n",
            "loss: 0.871022  [10880/28709]\n",
            "loss: 0.909601  [11520/28709]\n",
            "loss: 0.911689  [12160/28709]\n",
            "loss: 0.675988  [12800/28709]\n",
            "loss: 1.182626  [13440/28709]\n",
            "loss: 1.075619  [14080/28709]\n",
            "loss: 0.907735  [14720/28709]\n",
            "loss: 0.761870  [15360/28709]\n",
            "loss: 1.003134  [16000/28709]\n",
            "loss: 0.725319  [16640/28709]\n",
            "loss: 0.930857  [17280/28709]\n",
            "loss: 0.963158  [17920/28709]\n",
            "loss: 0.851387  [18560/28709]\n",
            "loss: 0.894572  [19200/28709]\n",
            "loss: 0.775226  [19840/28709]\n",
            "loss: 0.690737  [20480/28709]\n",
            "loss: 0.743447  [21120/28709]\n",
            "loss: 0.684264  [21760/28709]\n",
            "loss: 0.870303  [22400/28709]\n",
            "loss: 0.983144  [23040/28709]\n",
            "loss: 0.810203  [23680/28709]\n",
            "loss: 0.753268  [24320/28709]\n",
            "loss: 0.879665  [24960/28709]\n",
            "loss: 0.745212  [25600/28709]\n",
            "loss: 0.871793  [26240/28709]\n",
            "loss: 0.861353  [26880/28709]\n",
            "loss: 1.109737  [27520/28709]\n",
            "loss: 0.935368  [28160/28709]\n",
            "\n",
            "Test Error\n",
            "\tAccuracy: 58.3%, Avg loss: 1.244892 \n",
            "\n",
            "Epoch 32\n",
            "-------------------------------\n",
            "loss: 0.956184  [    0/28709]\n",
            "loss: 0.894082  [  640/28709]\n",
            "loss: 1.051131  [ 1280/28709]\n",
            "loss: 0.821019  [ 1920/28709]\n",
            "loss: 0.852720  [ 2560/28709]\n",
            "loss: 0.921748  [ 3200/28709]\n",
            "loss: 0.914608  [ 3840/28709]\n",
            "loss: 0.557194  [ 4480/28709]\n",
            "loss: 0.838487  [ 5120/28709]\n",
            "loss: 0.626487  [ 5760/28709]\n",
            "loss: 0.925653  [ 6400/28709]\n",
            "loss: 0.873972  [ 7040/28709]\n",
            "loss: 0.796644  [ 7680/28709]\n",
            "loss: 0.688182  [ 8320/28709]\n",
            "loss: 0.948095  [ 8960/28709]\n",
            "loss: 0.635730  [ 9600/28709]\n",
            "loss: 0.899679  [10240/28709]\n",
            "loss: 0.977215  [10880/28709]\n",
            "loss: 0.695975  [11520/28709]\n",
            "loss: 0.695814  [12160/28709]\n",
            "loss: 0.756514  [12800/28709]\n",
            "loss: 0.789925  [13440/28709]\n",
            "loss: 0.783089  [14080/28709]\n",
            "loss: 0.755618  [14720/28709]\n",
            "loss: 1.184061  [15360/28709]\n",
            "loss: 0.734436  [16000/28709]\n",
            "loss: 0.623881  [16640/28709]\n",
            "loss: 1.017750  [17280/28709]\n",
            "loss: 0.876992  [17920/28709]\n",
            "loss: 0.912651  [18560/28709]\n",
            "loss: 0.887045  [19200/28709]\n",
            "loss: 0.883522  [19840/28709]\n",
            "loss: 0.769363  [20480/28709]\n",
            "loss: 0.948210  [21120/28709]\n",
            "loss: 0.721672  [21760/28709]\n",
            "loss: 0.937556  [22400/28709]\n",
            "loss: 0.958193  [23040/28709]\n",
            "loss: 1.016095  [23680/28709]\n",
            "loss: 0.785747  [24320/28709]\n",
            "loss: 0.755689  [24960/28709]\n",
            "loss: 1.072411  [25600/28709]\n",
            "loss: 0.802766  [26240/28709]\n",
            "loss: 0.677556  [26880/28709]\n",
            "loss: 0.944557  [27520/28709]\n",
            "loss: 0.644532  [28160/28709]\n",
            "\n",
            "Test Error\n",
            "\tAccuracy: 57.8%, Avg loss: 1.216383 \n",
            "\n",
            "Epoch 33\n",
            "-------------------------------\n",
            "loss: 0.891258  [    0/28709]\n",
            "loss: 0.740741  [  640/28709]\n",
            "loss: 0.626770  [ 1280/28709]\n",
            "loss: 0.805324  [ 1920/28709]\n",
            "loss: 0.602779  [ 2560/28709]\n",
            "loss: 0.771897  [ 3200/28709]\n",
            "loss: 0.779603  [ 3840/28709]\n",
            "loss: 0.765126  [ 4480/28709]\n",
            "loss: 0.835428  [ 5120/28709]\n",
            "loss: 0.695925  [ 5760/28709]\n",
            "loss: 0.902327  [ 6400/28709]\n",
            "loss: 0.730497  [ 7040/28709]\n",
            "loss: 0.903212  [ 7680/28709]\n",
            "loss: 0.701333  [ 8320/28709]\n",
            "loss: 0.797123  [ 8960/28709]\n",
            "loss: 0.852263  [ 9600/28709]\n",
            "loss: 1.100606  [10240/28709]\n",
            "loss: 0.918278  [10880/28709]\n",
            "loss: 0.957306  [11520/28709]\n",
            "loss: 0.918538  [12160/28709]\n",
            "loss: 0.955899  [12800/28709]\n",
            "loss: 0.887314  [13440/28709]\n",
            "loss: 0.710633  [14080/28709]\n",
            "loss: 0.617993  [14720/28709]\n",
            "loss: 0.921938  [15360/28709]\n",
            "loss: 0.762135  [16000/28709]\n",
            "loss: 0.680741  [16640/28709]\n",
            "loss: 0.906865  [17280/28709]\n",
            "loss: 0.838695  [17920/28709]\n",
            "loss: 0.909022  [18560/28709]\n",
            "loss: 0.929209  [19200/28709]\n",
            "loss: 0.822952  [19840/28709]\n",
            "loss: 0.899886  [20480/28709]\n",
            "loss: 0.755101  [21120/28709]\n",
            "loss: 0.951160  [21760/28709]\n",
            "loss: 0.858058  [22400/28709]\n",
            "loss: 0.958767  [23040/28709]\n",
            "loss: 0.888231  [23680/28709]\n",
            "loss: 0.741144  [24320/28709]\n",
            "loss: 0.728531  [24960/28709]\n",
            "loss: 1.031383  [25600/28709]\n",
            "loss: 0.838491  [26240/28709]\n",
            "loss: 1.089870  [26880/28709]\n",
            "loss: 0.923496  [27520/28709]\n",
            "loss: 0.828954  [28160/28709]\n",
            "\n",
            "Test Error\n",
            "\tAccuracy: 50.8%, Avg loss: 2.866039 \n",
            "\n",
            "Epoch 34\n",
            "-------------------------------\n",
            "loss: 0.899250  [    0/28709]\n",
            "loss: 1.107321  [  640/28709]\n",
            "loss: 0.958170  [ 1280/28709]\n",
            "loss: 0.991606  [ 1920/28709]\n",
            "loss: 0.990519  [ 2560/28709]\n",
            "loss: 0.848405  [ 3200/28709]\n",
            "loss: 0.985867  [ 3840/28709]\n",
            "loss: 0.993096  [ 4480/28709]\n",
            "loss: 0.922593  [ 5120/28709]\n",
            "loss: 1.365546  [ 5760/28709]\n",
            "loss: 0.858782  [ 6400/28709]\n",
            "loss: 0.956616  [ 7040/28709]\n",
            "loss: 0.825012  [ 7680/28709]\n",
            "loss: 0.933158  [ 8320/28709]\n",
            "loss: 0.883902  [ 8960/28709]\n",
            "loss: 0.759018  [ 9600/28709]\n",
            "loss: 0.979276  [10240/28709]\n",
            "loss: 0.906966  [10880/28709]\n",
            "loss: 0.706710  [11520/28709]\n",
            "loss: 0.807358  [12160/28709]\n",
            "loss: 0.937499  [12800/28709]\n",
            "loss: 0.783091  [13440/28709]\n",
            "loss: 0.831251  [14080/28709]\n",
            "loss: 0.885978  [14720/28709]\n",
            "loss: 0.994374  [15360/28709]\n",
            "loss: 1.074877  [16000/28709]\n",
            "loss: 0.861503  [16640/28709]\n",
            "loss: 0.880659  [17280/28709]\n",
            "loss: 0.821727  [17920/28709]\n",
            "loss: 0.904596  [18560/28709]\n",
            "loss: 0.670746  [19200/28709]\n",
            "loss: 1.144774  [19840/28709]\n",
            "loss: 0.728375  [20480/28709]\n",
            "loss: 0.997697  [21120/28709]\n",
            "loss: 1.051410  [21760/28709]\n",
            "loss: 0.733951  [22400/28709]\n",
            "loss: 0.816283  [23040/28709]\n",
            "loss: 0.834915  [23680/28709]\n",
            "loss: 0.894848  [24320/28709]\n",
            "loss: 0.703718  [24960/28709]\n",
            "loss: 0.913709  [25600/28709]\n",
            "loss: 0.675265  [26240/28709]\n",
            "loss: 0.899871  [26880/28709]\n",
            "loss: 0.839356  [27520/28709]\n",
            "loss: 0.770856  [28160/28709]\n",
            "\n",
            "Test Error\n",
            "\tAccuracy: 57.3%, Avg loss: 1.189391 \n",
            "\n",
            "Epoch 35\n",
            "-------------------------------\n",
            "loss: 0.815080  [    0/28709]\n",
            "loss: 0.799167  [  640/28709]\n",
            "loss: 0.777171  [ 1280/28709]\n",
            "loss: 0.666990  [ 1920/28709]\n",
            "loss: 0.820001  [ 2560/28709]\n",
            "loss: 0.648596  [ 3200/28709]\n",
            "loss: 0.858264  [ 3840/28709]\n",
            "loss: 0.839809  [ 4480/28709]\n",
            "loss: 0.772931  [ 5120/28709]\n",
            "loss: 0.803813  [ 5760/28709]\n",
            "loss: 0.891984  [ 6400/28709]\n",
            "loss: 0.714797  [ 7040/28709]\n",
            "loss: 0.680893  [ 7680/28709]\n",
            "loss: 0.795094  [ 8320/28709]\n",
            "loss: 0.649626  [ 8960/28709]\n",
            "loss: 0.847289  [ 9600/28709]\n",
            "loss: 0.718283  [10240/28709]\n",
            "loss: 0.771897  [10880/28709]\n",
            "loss: 0.717105  [11520/28709]\n",
            "loss: 0.782739  [12160/28709]\n",
            "loss: 0.871139  [12800/28709]\n",
            "loss: 0.625402  [13440/28709]\n",
            "loss: 0.822464  [14080/28709]\n",
            "loss: 0.821586  [14720/28709]\n",
            "loss: 0.773187  [15360/28709]\n",
            "loss: 0.822636  [16000/28709]\n",
            "loss: 0.688014  [16640/28709]\n",
            "loss: 0.827197  [17280/28709]\n",
            "loss: 0.842275  [17920/28709]\n",
            "loss: 0.755180  [18560/28709]\n",
            "loss: 0.719471  [19200/28709]\n",
            "loss: 0.818357  [19840/28709]\n",
            "loss: 0.604707  [20480/28709]\n",
            "loss: 0.709001  [21120/28709]\n",
            "loss: 0.786658  [21760/28709]\n",
            "loss: 0.990479  [22400/28709]\n",
            "loss: 0.839270  [23040/28709]\n",
            "loss: 0.717249  [23680/28709]\n",
            "loss: 0.580848  [24320/28709]\n",
            "loss: 0.701294  [24960/28709]\n",
            "loss: 1.002505  [25600/28709]\n",
            "loss: 0.792468  [26240/28709]\n",
            "loss: 0.536364  [26880/28709]\n",
            "loss: 1.000311  [27520/28709]\n",
            "loss: 0.838996  [28160/28709]\n",
            "\n",
            "Test Error\n",
            "\tAccuracy: 58.3%, Avg loss: 1.182967 \n",
            "\n",
            "Epoch 36\n",
            "-------------------------------\n",
            "loss: 0.752667  [    0/28709]\n",
            "loss: 0.773053  [  640/28709]\n",
            "loss: 0.724488  [ 1280/28709]\n",
            "loss: 0.813526  [ 1920/28709]\n",
            "loss: 0.739274  [ 2560/28709]\n",
            "loss: 0.692185  [ 3200/28709]\n",
            "loss: 0.779666  [ 3840/28709]\n",
            "loss: 0.634753  [ 4480/28709]\n",
            "loss: 0.766786  [ 5120/28709]\n",
            "loss: 0.779813  [ 5760/28709]\n",
            "loss: 0.838874  [ 6400/28709]\n",
            "loss: 0.874387  [ 7040/28709]\n",
            "loss: 0.891794  [ 7680/28709]\n",
            "loss: 0.752850  [ 8320/28709]\n",
            "loss: 0.895652  [ 8960/28709]\n",
            "loss: 0.851904  [ 9600/28709]\n",
            "loss: 0.711172  [10240/28709]\n",
            "loss: 0.769832  [10880/28709]\n",
            "loss: 1.012423  [11520/28709]\n",
            "loss: 0.708517  [12160/28709]\n",
            "loss: 0.795348  [12800/28709]\n",
            "loss: 0.926594  [13440/28709]\n",
            "loss: 0.733502  [14080/28709]\n",
            "loss: 0.731887  [14720/28709]\n",
            "loss: 0.660007  [15360/28709]\n",
            "loss: 0.799057  [16000/28709]\n",
            "loss: 0.975887  [16640/28709]\n",
            "loss: 0.958392  [17280/28709]\n",
            "loss: 0.883599  [17920/28709]\n",
            "loss: 0.931199  [18560/28709]\n",
            "loss: 0.903435  [19200/28709]\n",
            "loss: 0.852187  [19840/28709]\n",
            "loss: 0.761191  [20480/28709]\n",
            "loss: 0.784111  [21120/28709]\n",
            "loss: 0.791034  [21760/28709]\n",
            "loss: 0.938884  [22400/28709]\n",
            "loss: 0.968481  [23040/28709]\n",
            "loss: 0.638956  [23680/28709]\n",
            "loss: 0.820596  [24320/28709]\n",
            "loss: 0.677589  [24960/28709]\n",
            "loss: 1.016204  [25600/28709]\n",
            "loss: 0.904389  [26240/28709]\n",
            "loss: 0.649818  [26880/28709]\n",
            "loss: 0.952029  [27520/28709]\n",
            "loss: 0.743816  [28160/28709]\n",
            "\n",
            "Test Error\n",
            "\tAccuracy: 57.0%, Avg loss: 1.199156 \n",
            "\n",
            "Epoch 37\n",
            "-------------------------------\n",
            "loss: 0.823438  [    0/28709]\n",
            "loss: 0.895184  [  640/28709]\n",
            "loss: 1.026094  [ 1280/28709]\n",
            "loss: 1.005238  [ 1920/28709]\n",
            "loss: 0.923172  [ 2560/28709]\n",
            "loss: 0.897471  [ 3200/28709]\n",
            "loss: 0.987882  [ 3840/28709]\n",
            "loss: 1.168783  [ 4480/28709]\n",
            "loss: 0.620271  [ 5120/28709]\n",
            "loss: 1.005539  [ 5760/28709]\n",
            "loss: 0.734010  [ 6400/28709]\n",
            "loss: 0.748750  [ 7040/28709]\n",
            "loss: 0.835218  [ 7680/28709]\n",
            "loss: 0.684994  [ 8320/28709]\n",
            "loss: 0.770951  [ 8960/28709]\n",
            "loss: 0.831103  [ 9600/28709]\n",
            "loss: 0.626509  [10240/28709]\n",
            "loss: 0.610123  [10880/28709]\n",
            "loss: 0.782257  [11520/28709]\n",
            "loss: 0.959754  [12160/28709]\n",
            "loss: 0.696753  [12800/28709]\n",
            "loss: 0.749530  [13440/28709]\n",
            "loss: 0.838854  [14080/28709]\n",
            "loss: 1.022660  [14720/28709]\n",
            "loss: 0.754336  [15360/28709]\n",
            "loss: 1.011330  [16000/28709]\n",
            "loss: 0.778125  [16640/28709]\n",
            "loss: 0.707987  [17280/28709]\n",
            "loss: 0.771939  [17920/28709]\n",
            "loss: 0.689744  [18560/28709]\n",
            "loss: 0.855774  [19200/28709]\n",
            "loss: 0.801157  [19840/28709]\n",
            "loss: 0.747357  [20480/28709]\n",
            "loss: 0.787100  [21120/28709]\n",
            "loss: 0.830491  [21760/28709]\n",
            "loss: 0.764551  [22400/28709]\n",
            "loss: 0.975218  [23040/28709]\n",
            "loss: 0.723829  [23680/28709]\n",
            "loss: 0.805839  [24320/28709]\n",
            "loss: 0.792062  [24960/28709]\n",
            "loss: 0.764492  [25600/28709]\n",
            "loss: 0.765651  [26240/28709]\n",
            "loss: 0.810139  [26880/28709]\n",
            "loss: 1.022367  [27520/28709]\n",
            "loss: 0.908296  [28160/28709]\n",
            "\n",
            "Test Error\n",
            "\tAccuracy: 57.5%, Avg loss: 1.211463 \n",
            "\n",
            "Epoch 38\n",
            "-------------------------------\n",
            "loss: 0.776366  [    0/28709]\n",
            "loss: 0.961676  [  640/28709]\n",
            "loss: 0.813504  [ 1280/28709]\n",
            "loss: 0.820963  [ 1920/28709]\n",
            "loss: 0.573525  [ 2560/28709]\n",
            "loss: 0.648405  [ 3200/28709]\n",
            "loss: 0.657160  [ 3840/28709]\n",
            "loss: 0.877231  [ 4480/28709]\n",
            "loss: 0.920352  [ 5120/28709]\n",
            "loss: 0.683686  [ 5760/28709]\n",
            "loss: 0.681304  [ 6400/28709]\n",
            "loss: 0.737048  [ 7040/28709]\n",
            "loss: 0.685366  [ 7680/28709]\n",
            "loss: 0.873943  [ 8320/28709]\n",
            "loss: 0.903874  [ 8960/28709]\n",
            "loss: 0.884443  [ 9600/28709]\n",
            "loss: 0.761254  [10240/28709]\n",
            "loss: 0.781963  [10880/28709]\n",
            "loss: 0.800042  [11520/28709]\n",
            "loss: 0.620751  [12160/28709]\n",
            "loss: 0.874909  [12800/28709]\n",
            "loss: 1.000554  [13440/28709]\n",
            "loss: 0.703054  [14080/28709]\n",
            "loss: 0.630291  [14720/28709]\n",
            "loss: 0.743259  [15360/28709]\n",
            "loss: 0.780345  [16000/28709]\n",
            "loss: 0.861094  [16640/28709]\n",
            "loss: 0.642847  [17280/28709]\n",
            "loss: 0.741777  [17920/28709]\n",
            "loss: 0.855440  [18560/28709]\n",
            "loss: 0.691893  [19200/28709]\n",
            "loss: 0.976969  [19840/28709]\n",
            "loss: 0.509365  [20480/28709]\n",
            "loss: 0.634311  [21120/28709]\n",
            "loss: 0.759957  [21760/28709]\n",
            "loss: 0.693304  [22400/28709]\n",
            "loss: 0.855403  [23040/28709]\n",
            "loss: 0.727538  [23680/28709]\n",
            "loss: 0.752321  [24320/28709]\n",
            "loss: 0.719431  [24960/28709]\n",
            "loss: 0.787095  [25600/28709]\n",
            "loss: 0.928948  [26240/28709]\n",
            "loss: 0.914404  [26880/28709]\n",
            "loss: 0.728836  [27520/28709]\n",
            "loss: 0.970808  [28160/28709]\n",
            "\n",
            "Test Error\n",
            "\tAccuracy: 58.4%, Avg loss: 1.185478 \n",
            "\n",
            "Epoch 39\n",
            "-------------------------------\n",
            "loss: 1.026745  [    0/28709]\n",
            "loss: 0.896564  [  640/28709]\n",
            "loss: 0.579217  [ 1280/28709]\n",
            "loss: 0.809864  [ 1920/28709]\n",
            "loss: 0.748426  [ 2560/28709]\n",
            "loss: 0.884007  [ 3200/28709]\n",
            "loss: 0.816039  [ 3840/28709]\n",
            "loss: 0.843125  [ 4480/28709]\n",
            "loss: 0.943820  [ 5120/28709]\n",
            "loss: 1.266992  [ 5760/28709]\n",
            "loss: 0.894932  [ 6400/28709]\n",
            "loss: 0.767270  [ 7040/28709]\n",
            "loss: 0.924308  [ 7680/28709]\n",
            "loss: 0.666879  [ 8320/28709]\n",
            "loss: 0.690158  [ 8960/28709]\n",
            "loss: 0.682589  [ 9600/28709]\n",
            "loss: 0.870656  [10240/28709]\n",
            "loss: 0.831733  [10880/28709]\n",
            "loss: 0.580880  [11520/28709]\n",
            "loss: 0.726192  [12160/28709]\n",
            "loss: 0.841859  [12800/28709]\n",
            "loss: 0.760822  [13440/28709]\n",
            "loss: 0.666772  [14080/28709]\n",
            "loss: 0.551586  [14720/28709]\n",
            "loss: 0.892756  [15360/28709]\n",
            "loss: 0.780646  [16000/28709]\n",
            "loss: 0.797854  [16640/28709]\n",
            "loss: 0.798829  [17280/28709]\n",
            "loss: 0.714392  [17920/28709]\n",
            "loss: 0.953644  [18560/28709]\n",
            "loss: 0.716936  [19200/28709]\n",
            "loss: 0.797222  [19840/28709]\n",
            "loss: 0.910596  [20480/28709]\n",
            "loss: 0.966607  [21120/28709]\n",
            "loss: 0.564443  [21760/28709]\n",
            "loss: 1.123139  [22400/28709]\n",
            "loss: 0.947183  [23040/28709]\n",
            "loss: 1.014050  [23680/28709]\n",
            "loss: 0.526556  [24320/28709]\n",
            "loss: 0.742888  [24960/28709]\n",
            "loss: 0.907337  [25600/28709]\n",
            "loss: 0.689228  [26240/28709]\n",
            "loss: 0.662954  [26880/28709]\n",
            "loss: 0.912229  [27520/28709]\n",
            "loss: 0.659584  [28160/28709]\n",
            "\n",
            "Test Error\n",
            "\tAccuracy: 58.1%, Avg loss: 1.247930 \n",
            "\n",
            "Epoch 40\n",
            "-------------------------------\n",
            "loss: 0.855995  [    0/28709]\n",
            "loss: 0.802077  [  640/28709]\n",
            "loss: 1.082182  [ 1280/28709]\n",
            "loss: 0.712477  [ 1920/28709]\n",
            "loss: 1.050652  [ 2560/28709]\n",
            "loss: 0.683095  [ 3200/28709]\n",
            "loss: 0.742432  [ 3840/28709]\n",
            "loss: 0.821534  [ 4480/28709]\n",
            "loss: 0.703406  [ 5120/28709]\n",
            "loss: 0.545397  [ 5760/28709]\n",
            "loss: 0.698338  [ 6400/28709]\n",
            "loss: 0.707116  [ 7040/28709]\n",
            "loss: 0.832999  [ 7680/28709]\n",
            "loss: 0.554481  [ 8320/28709]\n",
            "loss: 0.702398  [ 8960/28709]\n",
            "loss: 0.515732  [ 9600/28709]\n",
            "loss: 0.720385  [10240/28709]\n",
            "loss: 1.095758  [10880/28709]\n",
            "loss: 0.665429  [11520/28709]\n",
            "loss: 0.669496  [12160/28709]\n",
            "loss: 0.723687  [12800/28709]\n",
            "loss: 0.662490  [13440/28709]\n",
            "loss: 0.660341  [14080/28709]\n",
            "loss: 0.941756  [14720/28709]\n",
            "loss: 0.906621  [15360/28709]\n",
            "loss: 0.650003  [16000/28709]\n",
            "loss: 0.698694  [16640/28709]\n",
            "loss: 0.878434  [17280/28709]\n",
            "loss: 0.878353  [17920/28709]\n",
            "loss: 0.574096  [18560/28709]\n",
            "loss: 0.824529  [19200/28709]\n",
            "loss: 0.722243  [19840/28709]\n",
            "loss: 0.683610  [20480/28709]\n",
            "loss: 0.819437  [21120/28709]\n",
            "loss: 0.896485  [21760/28709]\n",
            "loss: 0.759199  [22400/28709]\n",
            "loss: 1.124740  [23040/28709]\n",
            "loss: 0.748451  [23680/28709]\n",
            "loss: 0.757392  [24320/28709]\n",
            "loss: 0.694454  [24960/28709]\n",
            "loss: 0.657877  [25600/28709]\n",
            "loss: 0.787132  [26240/28709]\n",
            "loss: 0.624945  [26880/28709]\n",
            "loss: 0.847765  [27520/28709]\n",
            "loss: 0.799623  [28160/28709]\n",
            "\n",
            "Test Error\n",
            "\tAccuracy: 58.0%, Avg loss: 1.278507 \n",
            "\n",
            "Epoch 41\n",
            "-------------------------------\n",
            "loss: 0.731633  [    0/28709]\n",
            "loss: 0.809938  [  640/28709]\n",
            "loss: 0.580496  [ 1280/28709]\n",
            "loss: 0.693706  [ 1920/28709]\n",
            "loss: 0.651783  [ 2560/28709]\n",
            "loss: 0.806793  [ 3200/28709]\n",
            "loss: 0.681973  [ 3840/28709]\n",
            "loss: 0.497140  [ 4480/28709]\n",
            "loss: 0.565406  [ 5120/28709]\n",
            "loss: 0.512511  [ 5760/28709]\n",
            "loss: 0.880678  [ 6400/28709]\n",
            "loss: 0.656337  [ 7040/28709]\n",
            "loss: 0.898224  [ 7680/28709]\n",
            "loss: 0.806011  [ 8320/28709]\n",
            "loss: 0.609963  [ 8960/28709]\n",
            "loss: 0.775692  [ 9600/28709]\n",
            "loss: 0.716043  [10240/28709]\n",
            "loss: 1.261098  [10880/28709]\n",
            "loss: 0.992451  [11520/28709]\n",
            "loss: 0.770645  [12160/28709]\n",
            "loss: 0.740146  [12800/28709]\n",
            "loss: 0.662229  [13440/28709]\n",
            "loss: 0.651621  [14080/28709]\n",
            "loss: 0.852486  [14720/28709]\n",
            "loss: 0.755096  [15360/28709]\n",
            "loss: 0.614633  [16000/28709]\n",
            "loss: 0.860656  [16640/28709]\n",
            "loss: 0.825307  [17280/28709]\n",
            "loss: 0.508813  [17920/28709]\n",
            "loss: 0.720686  [18560/28709]\n",
            "loss: 0.769218  [19200/28709]\n",
            "loss: 0.902460  [19840/28709]\n",
            "loss: 0.878930  [20480/28709]\n",
            "loss: 0.864922  [21120/28709]\n",
            "loss: 0.687521  [21760/28709]\n",
            "loss: 0.573727  [22400/28709]\n",
            "loss: 0.857062  [23040/28709]\n",
            "loss: 0.630672  [23680/28709]\n",
            "loss: 0.690403  [24320/28709]\n",
            "loss: 0.592893  [24960/28709]\n",
            "loss: 0.909333  [25600/28709]\n",
            "loss: 0.640935  [26240/28709]\n",
            "loss: 0.683819  [26880/28709]\n",
            "loss: 0.640381  [27520/28709]\n",
            "loss: 0.576004  [28160/28709]\n",
            "\n",
            "Test Error\n",
            "\tAccuracy: 58.3%, Avg loss: 1.251413 \n",
            "\n",
            "Epoch 42\n",
            "-------------------------------\n",
            "loss: 0.803939  [    0/28709]\n",
            "loss: 0.863958  [  640/28709]\n",
            "loss: 0.799632  [ 1280/28709]\n",
            "loss: 0.719934  [ 1920/28709]\n",
            "loss: 0.706533  [ 2560/28709]\n",
            "loss: 0.638707  [ 3200/28709]\n",
            "loss: 0.548370  [ 3840/28709]\n",
            "loss: 0.606020  [ 4480/28709]\n",
            "loss: 0.652031  [ 5120/28709]\n",
            "loss: 0.823084  [ 5760/28709]\n",
            "loss: 0.800379  [ 6400/28709]\n",
            "loss: 0.681302  [ 7040/28709]\n",
            "loss: 0.586359  [ 7680/28709]\n",
            "loss: 0.706323  [ 8320/28709]\n",
            "loss: 0.823354  [ 8960/28709]\n",
            "loss: 0.913733  [ 9600/28709]\n",
            "loss: 0.797725  [10240/28709]\n",
            "loss: 0.842716  [10880/28709]\n",
            "loss: 0.657555  [11520/28709]\n",
            "loss: 0.710909  [12160/28709]\n",
            "loss: 0.858271  [12800/28709]\n",
            "loss: 0.711208  [13440/28709]\n",
            "loss: 0.750369  [14080/28709]\n",
            "loss: 0.809609  [14720/28709]\n",
            "loss: 0.706565  [15360/28709]\n",
            "loss: 0.887572  [16000/28709]\n",
            "loss: 0.628810  [16640/28709]\n",
            "loss: 0.655197  [17280/28709]\n",
            "loss: 0.920495  [17920/28709]\n",
            "loss: 0.557642  [18560/28709]\n",
            "loss: 0.688920  [19200/28709]\n",
            "loss: 0.711193  [19840/28709]\n",
            "loss: 0.697598  [20480/28709]\n",
            "loss: 0.629210  [21120/28709]\n",
            "loss: 0.905368  [21760/28709]\n",
            "loss: 0.608963  [22400/28709]\n",
            "loss: 1.029105  [23040/28709]\n",
            "loss: 1.142016  [23680/28709]\n",
            "loss: 1.113682  [24320/28709]\n",
            "loss: 0.966656  [24960/28709]\n",
            "loss: 0.873270  [25600/28709]\n",
            "loss: 0.762083  [26240/28709]\n",
            "loss: 0.863743  [26880/28709]\n",
            "loss: 0.680533  [27520/28709]\n",
            "loss: 0.901547  [28160/28709]\n",
            "\n",
            "Test Error\n",
            "\tAccuracy: 58.0%, Avg loss: 1.239620 \n",
            "\n",
            "Epoch 43\n",
            "-------------------------------\n",
            "loss: 0.750428  [    0/28709]\n",
            "loss: 0.866031  [  640/28709]\n",
            "loss: 0.656309  [ 1280/28709]\n",
            "loss: 0.523087  [ 1920/28709]\n",
            "loss: 0.958671  [ 2560/28709]\n",
            "loss: 0.798427  [ 3200/28709]\n",
            "loss: 0.794325  [ 3840/28709]\n",
            "loss: 0.832218  [ 4480/28709]\n",
            "loss: 0.957514  [ 5120/28709]\n",
            "loss: 0.753632  [ 5760/28709]\n",
            "loss: 0.951306  [ 6400/28709]\n",
            "loss: 0.896730  [ 7040/28709]\n",
            "loss: 1.003914  [ 7680/28709]\n",
            "loss: 0.691769  [ 8320/28709]\n",
            "loss: 0.769688  [ 8960/28709]\n",
            "loss: 0.664942  [ 9600/28709]\n",
            "loss: 0.880440  [10240/28709]\n",
            "loss: 0.839027  [10880/28709]\n",
            "loss: 0.866627  [11520/28709]\n",
            "loss: 0.825055  [12160/28709]\n",
            "loss: 1.165188  [12800/28709]\n",
            "loss: 0.942284  [13440/28709]\n",
            "loss: 0.727162  [14080/28709]\n",
            "loss: 0.658983  [14720/28709]\n",
            "loss: 0.890810  [15360/28709]\n",
            "loss: 0.643743  [16000/28709]\n",
            "loss: 0.608581  [16640/28709]\n",
            "loss: 0.738606  [17280/28709]\n",
            "loss: 0.655526  [17920/28709]\n",
            "loss: 0.761003  [18560/28709]\n",
            "loss: 0.906028  [19200/28709]\n",
            "loss: 0.959861  [19840/28709]\n",
            "loss: 0.821592  [20480/28709]\n",
            "loss: 0.717772  [21120/28709]\n",
            "loss: 0.840022  [21760/28709]\n",
            "loss: 0.644788  [22400/28709]\n",
            "loss: 0.553271  [23040/28709]\n",
            "loss: 0.792459  [23680/28709]\n",
            "loss: 0.495337  [24320/28709]\n",
            "loss: 0.541511  [24960/28709]\n",
            "loss: 0.668320  [25600/28709]\n",
            "loss: 0.832113  [26240/28709]\n",
            "loss: 0.846383  [26880/28709]\n",
            "loss: 0.829610  [27520/28709]\n",
            "loss: 0.596024  [28160/28709]\n",
            "\n",
            "Test Error\n",
            "\tAccuracy: 58.1%, Avg loss: 1.558090 \n",
            "\n",
            "Epoch 44\n",
            "-------------------------------\n",
            "loss: 0.776000  [    0/28709]\n",
            "loss: 0.535009  [  640/28709]\n",
            "loss: 0.613690  [ 1280/28709]\n",
            "loss: 0.851345  [ 1920/28709]\n",
            "loss: 0.668166  [ 2560/28709]\n",
            "loss: 0.834700  [ 3200/28709]\n",
            "loss: 0.677735  [ 3840/28709]\n",
            "loss: 0.793059  [ 4480/28709]\n",
            "loss: 0.715319  [ 5120/28709]\n",
            "loss: 0.544838  [ 5760/28709]\n",
            "loss: 0.748481  [ 6400/28709]\n",
            "loss: 0.780712  [ 7040/28709]\n",
            "loss: 0.653318  [ 7680/28709]\n",
            "loss: 0.720099  [ 8320/28709]\n",
            "loss: 0.620040  [ 8960/28709]\n",
            "loss: 0.673521  [ 9600/28709]\n",
            "loss: 0.742162  [10240/28709]\n",
            "loss: 0.605685  [10880/28709]\n",
            "loss: 1.339133  [11520/28709]\n",
            "loss: 0.925451  [12160/28709]\n",
            "loss: 0.928784  [12800/28709]\n",
            "loss: 1.011538  [13440/28709]\n",
            "loss: 0.667623  [14080/28709]\n",
            "loss: 0.959219  [14720/28709]\n",
            "loss: 0.735228  [15360/28709]\n",
            "loss: 0.779760  [16000/28709]\n",
            "loss: 0.686199  [16640/28709]\n",
            "loss: 0.728062  [17280/28709]\n",
            "loss: 0.736090  [17920/28709]\n",
            "loss: 0.912405  [18560/28709]\n",
            "loss: 1.019386  [19200/28709]\n",
            "loss: 0.782975  [19840/28709]\n",
            "loss: 0.712389  [20480/28709]\n",
            "loss: 0.759266  [21120/28709]\n",
            "loss: 0.754790  [21760/28709]\n",
            "loss: 0.606409  [22400/28709]\n",
            "loss: 0.864630  [23040/28709]\n",
            "loss: 0.779209  [23680/28709]\n",
            "loss: 0.794973  [24320/28709]\n",
            "loss: 0.684242  [24960/28709]\n",
            "loss: 0.836879  [25600/28709]\n",
            "loss: 0.853299  [26240/28709]\n",
            "loss: 0.758752  [26880/28709]\n",
            "loss: 0.724911  [27520/28709]\n",
            "loss: 0.514598  [28160/28709]\n",
            "\n",
            "Test Error\n",
            "\tAccuracy: 58.0%, Avg loss: 1.214422 \n",
            "\n",
            "Epoch 45\n",
            "-------------------------------\n",
            "loss: 0.813997  [    0/28709]\n",
            "loss: 0.681222  [  640/28709]\n",
            "loss: 0.723948  [ 1280/28709]\n",
            "loss: 0.877340  [ 1920/28709]\n",
            "loss: 0.646935  [ 2560/28709]\n",
            "loss: 0.879147  [ 3200/28709]\n",
            "loss: 0.687201  [ 3840/28709]\n",
            "loss: 0.824608  [ 4480/28709]\n",
            "loss: 0.768858  [ 5120/28709]\n",
            "loss: 0.523360  [ 5760/28709]\n",
            "loss: 0.741770  [ 6400/28709]\n",
            "loss: 0.761063  [ 7040/28709]\n",
            "loss: 0.674048  [ 7680/28709]\n",
            "loss: 0.616620  [ 8320/28709]\n",
            "loss: 0.791645  [ 8960/28709]\n",
            "loss: 0.761043  [ 9600/28709]\n",
            "loss: 0.677554  [10240/28709]\n",
            "loss: 0.510740  [10880/28709]\n",
            "loss: 0.603616  [11520/28709]\n",
            "loss: 0.768295  [12160/28709]\n",
            "loss: 0.793884  [12800/28709]\n",
            "loss: 0.923493  [13440/28709]\n",
            "loss: 0.675375  [14080/28709]\n",
            "loss: 0.706909  [14720/28709]\n",
            "loss: 0.566499  [15360/28709]\n",
            "loss: 0.804722  [16000/28709]\n",
            "loss: 0.821472  [16640/28709]\n",
            "loss: 0.803410  [17280/28709]\n",
            "loss: 0.816763  [17920/28709]\n",
            "loss: 0.866630  [18560/28709]\n",
            "loss: 0.774002  [19200/28709]\n",
            "loss: 0.725762  [19840/28709]\n",
            "loss: 0.612480  [20480/28709]\n",
            "loss: 0.653672  [21120/28709]\n",
            "loss: 0.748190  [21760/28709]\n",
            "loss: 0.748615  [22400/28709]\n",
            "loss: 0.773497  [23040/28709]\n",
            "loss: 0.684546  [23680/28709]\n",
            "loss: 0.925341  [24320/28709]\n",
            "loss: 0.638425  [24960/28709]\n",
            "loss: 0.735073  [25600/28709]\n",
            "loss: 0.724195  [26240/28709]\n",
            "loss: 0.743382  [26880/28709]\n",
            "loss: 0.735086  [27520/28709]\n",
            "loss: 0.460226  [28160/28709]\n",
            "\n",
            "Test Error\n",
            "\tAccuracy: 58.5%, Avg loss: 1.248980 \n",
            "\n",
            "Epoch 46\n",
            "-------------------------------\n",
            "loss: 0.696253  [    0/28709]\n",
            "loss: 0.742128  [  640/28709]\n",
            "loss: 0.925627  [ 1280/28709]\n",
            "loss: 0.551115  [ 1920/28709]\n",
            "loss: 0.729523  [ 2560/28709]\n",
            "loss: 0.635578  [ 3200/28709]\n",
            "loss: 0.541827  [ 3840/28709]\n",
            "loss: 0.653362  [ 4480/28709]\n",
            "loss: 0.841418  [ 5120/28709]\n",
            "loss: 0.906388  [ 5760/28709]\n",
            "loss: 0.505981  [ 6400/28709]\n",
            "loss: 0.540976  [ 7040/28709]\n",
            "loss: 0.793815  [ 7680/28709]\n",
            "loss: 0.621103  [ 8320/28709]\n",
            "loss: 0.744049  [ 8960/28709]\n",
            "loss: 0.734505  [ 9600/28709]\n",
            "loss: 0.752299  [10240/28709]\n",
            "loss: 0.709580  [10880/28709]\n",
            "loss: 0.558405  [11520/28709]\n",
            "loss: 0.920844  [12160/28709]\n",
            "loss: 0.536260  [12800/28709]\n",
            "loss: 0.733559  [13440/28709]\n",
            "loss: 0.565981  [14080/28709]\n",
            "loss: 0.672908  [14720/28709]\n",
            "loss: 0.883098  [15360/28709]\n",
            "loss: 0.687956  [16000/28709]\n",
            "loss: 0.588432  [16640/28709]\n",
            "loss: 0.445388  [17280/28709]\n",
            "loss: 0.784868  [17920/28709]\n",
            "loss: 0.788655  [18560/28709]\n",
            "loss: 0.879932  [19200/28709]\n",
            "loss: 0.741416  [19840/28709]\n",
            "loss: 0.778260  [20480/28709]\n",
            "loss: 0.677498  [21120/28709]\n",
            "loss: 0.671464  [21760/28709]\n",
            "loss: 0.713889  [22400/28709]\n",
            "loss: 0.533533  [23040/28709]\n",
            "loss: 0.571592  [23680/28709]\n",
            "loss: 0.709819  [24320/28709]\n",
            "loss: 0.864052  [24960/28709]\n",
            "loss: 0.690623  [25600/28709]\n",
            "loss: 0.698094  [26240/28709]\n",
            "loss: 0.884355  [26880/28709]\n",
            "loss: 0.862563  [27520/28709]\n",
            "loss: 0.828088  [28160/28709]\n",
            "\n",
            "Test Error\n",
            "\tAccuracy: 58.1%, Avg loss: 1.223705 \n",
            "\n",
            "Epoch 47\n",
            "-------------------------------\n",
            "loss: 0.782391  [    0/28709]\n",
            "loss: 0.935329  [  640/28709]\n",
            "loss: 0.746882  [ 1280/28709]\n",
            "loss: 0.564210  [ 1920/28709]\n",
            "loss: 0.759988  [ 2560/28709]\n",
            "loss: 0.795102  [ 3200/28709]\n",
            "loss: 0.606214  [ 3840/28709]\n",
            "loss: 0.721196  [ 4480/28709]\n",
            "loss: 0.777042  [ 5120/28709]\n",
            "loss: 0.592450  [ 5760/28709]\n",
            "loss: 0.851789  [ 6400/28709]\n",
            "loss: 0.763100  [ 7040/28709]\n",
            "loss: 0.743372  [ 7680/28709]\n",
            "loss: 0.744422  [ 8320/28709]\n",
            "loss: 1.064552  [ 8960/28709]\n",
            "loss: 0.743927  [ 9600/28709]\n",
            "loss: 0.753075  [10240/28709]\n",
            "loss: 0.700302  [10880/28709]\n",
            "loss: 0.782967  [11520/28709]\n",
            "loss: 0.812679  [12160/28709]\n",
            "loss: 0.826560  [12800/28709]\n",
            "loss: 0.909111  [13440/28709]\n",
            "loss: 0.827633  [14080/28709]\n",
            "loss: 0.587055  [14720/28709]\n",
            "loss: 0.621811  [15360/28709]\n",
            "loss: 0.753207  [16000/28709]\n",
            "loss: 0.719710  [16640/28709]\n",
            "loss: 0.809166  [17280/28709]\n",
            "loss: 0.674077  [17920/28709]\n",
            "loss: 1.052892  [18560/28709]\n",
            "loss: 0.468773  [19200/28709]\n",
            "loss: 0.578480  [19840/28709]\n",
            "loss: 0.675582  [20480/28709]\n",
            "loss: 0.949869  [21120/28709]\n",
            "loss: 0.903146  [21760/28709]\n",
            "loss: 0.706015  [22400/28709]\n",
            "loss: 0.579950  [23040/28709]\n",
            "loss: 0.603749  [23680/28709]\n",
            "loss: 0.756512  [24320/28709]\n",
            "loss: 0.620691  [24960/28709]\n",
            "loss: 0.832228  [25600/28709]\n",
            "loss: 0.685800  [26240/28709]\n",
            "loss: 0.796220  [26880/28709]\n",
            "loss: 0.772329  [27520/28709]\n",
            "loss: 0.489107  [28160/28709]\n",
            "\n",
            "Test Error\n",
            "\tAccuracy: 58.1%, Avg loss: 1.235162 \n",
            "\n",
            "Epoch 48\n",
            "-------------------------------\n",
            "loss: 0.600393  [    0/28709]\n",
            "loss: 0.760334  [  640/28709]\n",
            "loss: 0.826909  [ 1280/28709]\n",
            "loss: 0.543023  [ 1920/28709]\n",
            "loss: 0.552513  [ 2560/28709]\n",
            "loss: 0.725590  [ 3200/28709]\n",
            "loss: 0.600118  [ 3840/28709]\n",
            "loss: 0.892089  [ 4480/28709]\n",
            "loss: 0.646430  [ 5120/28709]\n",
            "loss: 0.599190  [ 5760/28709]\n",
            "loss: 0.729867  [ 6400/28709]\n",
            "loss: 0.568847  [ 7040/28709]\n",
            "loss: 0.623508  [ 7680/28709]\n",
            "loss: 0.748768  [ 8320/28709]\n",
            "loss: 0.608013  [ 8960/28709]\n",
            "loss: 0.529904  [ 9600/28709]\n",
            "loss: 0.477256  [10240/28709]\n",
            "loss: 0.436601  [10880/28709]\n",
            "loss: 0.586807  [11520/28709]\n",
            "loss: 0.795733  [12160/28709]\n",
            "loss: 0.643508  [12800/28709]\n",
            "loss: 0.493443  [13440/28709]\n",
            "loss: 0.857242  [14080/28709]\n",
            "loss: 0.466826  [14720/28709]\n",
            "loss: 0.619843  [15360/28709]\n",
            "loss: 0.508612  [16000/28709]\n",
            "loss: 0.998918  [16640/28709]\n",
            "loss: 0.630694  [17280/28709]\n",
            "loss: 0.603767  [17920/28709]\n",
            "loss: 0.808787  [18560/28709]\n",
            "loss: 0.806481  [19200/28709]\n",
            "loss: 0.774065  [19840/28709]\n",
            "loss: 0.943350  [20480/28709]\n",
            "loss: 0.718144  [21120/28709]\n",
            "loss: 0.764721  [21760/28709]\n",
            "loss: 0.657548  [22400/28709]\n",
            "loss: 0.708018  [23040/28709]\n",
            "loss: 0.813359  [23680/28709]\n",
            "loss: 0.719648  [24320/28709]\n",
            "loss: 0.762489  [24960/28709]\n",
            "loss: 0.719557  [25600/28709]\n",
            "loss: 0.476289  [26240/28709]\n",
            "loss: 0.775159  [26880/28709]\n",
            "loss: 0.531444  [27520/28709]\n",
            "loss: 0.938649  [28160/28709]\n",
            "\n",
            "Test Error\n",
            "\tAccuracy: 58.6%, Avg loss: 1.220920 \n",
            "\n",
            "Epoch 49\n",
            "-------------------------------\n",
            "loss: 0.650510  [    0/28709]\n",
            "loss: 0.617913  [  640/28709]\n",
            "loss: 0.449228  [ 1280/28709]\n",
            "loss: 0.709249  [ 1920/28709]\n",
            "loss: 0.706504  [ 2560/28709]\n",
            "loss: 0.639044  [ 3200/28709]\n",
            "loss: 0.751452  [ 3840/28709]\n",
            "loss: 0.706260  [ 4480/28709]\n",
            "loss: 0.856383  [ 5120/28709]\n",
            "loss: 0.715599  [ 5760/28709]\n",
            "loss: 0.701317  [ 6400/28709]\n",
            "loss: 0.580536  [ 7040/28709]\n",
            "loss: 0.710676  [ 7680/28709]\n",
            "loss: 0.620780  [ 8320/28709]\n",
            "loss: 0.573641  [ 8960/28709]\n",
            "loss: 0.790756  [ 9600/28709]\n",
            "loss: 0.760053  [10240/28709]\n",
            "loss: 0.638794  [10880/28709]\n",
            "loss: 0.708114  [11520/28709]\n",
            "loss: 0.785246  [12160/28709]\n",
            "loss: 0.782634  [12800/28709]\n",
            "loss: 0.603129  [13440/28709]\n",
            "loss: 0.815864  [14080/28709]\n",
            "loss: 0.651531  [14720/28709]\n",
            "loss: 0.659845  [15360/28709]\n",
            "loss: 0.702381  [16000/28709]\n",
            "loss: 0.698327  [16640/28709]\n",
            "loss: 0.723801  [17280/28709]\n",
            "loss: 0.554042  [17920/28709]\n",
            "loss: 0.669822  [18560/28709]\n",
            "loss: 0.621220  [19200/28709]\n",
            "loss: 0.678672  [19840/28709]\n",
            "loss: 0.637564  [20480/28709]\n",
            "loss: 0.864489  [21120/28709]\n",
            "loss: 0.678797  [21760/28709]\n",
            "loss: 0.560718  [22400/28709]\n",
            "loss: 0.713775  [23040/28709]\n",
            "loss: 0.786942  [23680/28709]\n",
            "loss: 0.637259  [24320/28709]\n",
            "loss: 0.892325  [24960/28709]\n",
            "loss: 0.726372  [25600/28709]\n",
            "loss: 0.542526  [26240/28709]\n",
            "loss: 0.663203  [26880/28709]\n",
            "loss: 0.665274  [27520/28709]\n",
            "loss: 0.899955  [28160/28709]\n",
            "\n",
            "Test Error\n",
            "\tAccuracy: 57.4%, Avg loss: 1.252475 \n",
            "\n",
            "Epoch 50\n",
            "-------------------------------\n",
            "loss: 0.588900  [    0/28709]\n",
            "loss: 0.569524  [  640/28709]\n",
            "loss: 0.547607  [ 1280/28709]\n",
            "loss: 0.665131  [ 1920/28709]\n",
            "loss: 0.606872  [ 2560/28709]\n",
            "loss: 0.562962  [ 3200/28709]\n",
            "loss: 0.824391  [ 3840/28709]\n",
            "loss: 0.618069  [ 4480/28709]\n",
            "loss: 0.572752  [ 5120/28709]\n",
            "loss: 0.946812  [ 5760/28709]\n",
            "loss: 0.689239  [ 6400/28709]\n",
            "loss: 0.814802  [ 7040/28709]\n",
            "loss: 0.730109  [ 7680/28709]\n",
            "loss: 0.918414  [ 8320/28709]\n",
            "loss: 0.460465  [ 8960/28709]\n",
            "loss: 0.795417  [ 9600/28709]\n",
            "loss: 1.003640  [10240/28709]\n",
            "loss: 0.769423  [10880/28709]\n",
            "loss: 0.755404  [11520/28709]\n",
            "loss: 0.479740  [12160/28709]\n",
            "loss: 0.585120  [12800/28709]\n",
            "loss: 0.761883  [13440/28709]\n",
            "loss: 0.656327  [14080/28709]\n",
            "loss: 0.496097  [14720/28709]\n",
            "loss: 0.832441  [15360/28709]\n",
            "loss: 0.676040  [16000/28709]\n",
            "loss: 0.570209  [16640/28709]\n",
            "loss: 0.437489  [17280/28709]\n",
            "loss: 0.682079  [17920/28709]\n",
            "loss: 0.656808  [18560/28709]\n",
            "loss: 0.774040  [19200/28709]\n",
            "loss: 0.710726  [19840/28709]\n",
            "loss: 0.793369  [20480/28709]\n",
            "loss: 0.780393  [21120/28709]\n",
            "loss: 0.615608  [21760/28709]\n",
            "loss: 0.861334  [22400/28709]\n",
            "loss: 0.865861  [23040/28709]\n",
            "loss: 0.703125  [23680/28709]\n",
            "loss: 0.680023  [24320/28709]\n",
            "loss: 0.705225  [24960/28709]\n",
            "loss: 0.790808  [25600/28709]\n",
            "loss: 0.963884  [26240/28709]\n",
            "loss: 0.636536  [26880/28709]\n",
            "loss: 0.688453  [27520/28709]\n",
            "loss: 0.421681  [28160/28709]\n",
            "\n",
            "Test Error\n",
            "\tAccuracy: 58.1%, Avg loss: 1.247771 \n",
            "\n",
            "Epoch 51\n",
            "-------------------------------\n",
            "loss: 0.505798  [    0/28709]\n",
            "loss: 0.652620  [  640/28709]\n",
            "loss: 0.750447  [ 1280/28709]\n",
            "loss: 0.612919  [ 1920/28709]\n",
            "loss: 0.551669  [ 2560/28709]\n",
            "loss: 0.606523  [ 3200/28709]\n",
            "loss: 0.621131  [ 3840/28709]\n",
            "loss: 0.576744  [ 4480/28709]\n",
            "loss: 0.871183  [ 5120/28709]\n",
            "loss: 0.611951  [ 5760/28709]\n",
            "loss: 0.544511  [ 6400/28709]\n",
            "loss: 0.980322  [ 7040/28709]\n",
            "loss: 0.555484  [ 7680/28709]\n",
            "loss: 0.536960  [ 8320/28709]\n",
            "loss: 0.668399  [ 8960/28709]\n",
            "loss: 0.569838  [ 9600/28709]\n",
            "loss: 0.685291  [10240/28709]\n",
            "loss: 0.640447  [10880/28709]\n",
            "loss: 0.728897  [11520/28709]\n",
            "loss: 0.500045  [12160/28709]\n",
            "loss: 0.665230  [12800/28709]\n",
            "loss: 0.617601  [13440/28709]\n",
            "loss: 0.536978  [14080/28709]\n",
            "loss: 0.616239  [14720/28709]\n",
            "loss: 0.699076  [15360/28709]\n",
            "loss: 0.485867  [16000/28709]\n",
            "loss: 0.704334  [16640/28709]\n",
            "loss: 0.807045  [17280/28709]\n",
            "loss: 0.636992  [17920/28709]\n",
            "loss: 0.850983  [18560/28709]\n",
            "loss: 0.565968  [19200/28709]\n",
            "loss: 0.792612  [19840/28709]\n",
            "loss: 0.587653  [20480/28709]\n",
            "loss: 0.697540  [21120/28709]\n",
            "loss: 0.731929  [21760/28709]\n",
            "loss: 0.605170  [22400/28709]\n",
            "loss: 0.640093  [23040/28709]\n",
            "loss: 0.585946  [23680/28709]\n",
            "loss: 0.630972  [24320/28709]\n",
            "loss: 0.612291  [24960/28709]\n",
            "loss: 0.795173  [25600/28709]\n",
            "loss: 0.596071  [26240/28709]\n",
            "loss: 0.667888  [26880/28709]\n",
            "loss: 0.511827  [27520/28709]\n",
            "loss: 0.597364  [28160/28709]\n",
            "\n",
            "Test Error\n",
            "\tAccuracy: 58.2%, Avg loss: 1.249912 \n",
            "\n",
            "Epoch 52\n",
            "-------------------------------\n",
            "loss: 0.751909  [    0/28709]\n",
            "loss: 0.733140  [  640/28709]\n",
            "loss: 0.562886  [ 1280/28709]\n",
            "loss: 0.537416  [ 1920/28709]\n",
            "loss: 0.611802  [ 2560/28709]\n",
            "loss: 0.606543  [ 3200/28709]\n",
            "loss: 0.693787  [ 3840/28709]\n",
            "loss: 0.848441  [ 4480/28709]\n",
            "loss: 0.590970  [ 5120/28709]\n",
            "loss: 0.559697  [ 5760/28709]\n",
            "loss: 0.668282  [ 6400/28709]\n",
            "loss: 0.738306  [ 7040/28709]\n",
            "loss: 0.664962  [ 7680/28709]\n",
            "loss: 0.565057  [ 8320/28709]\n",
            "loss: 0.676821  [ 8960/28709]\n",
            "loss: 0.487974  [ 9600/28709]\n",
            "loss: 0.710685  [10240/28709]\n",
            "loss: 0.628610  [10880/28709]\n",
            "loss: 0.682988  [11520/28709]\n",
            "loss: 0.714432  [12160/28709]\n",
            "loss: 0.655665  [12800/28709]\n",
            "loss: 0.546141  [13440/28709]\n",
            "loss: 0.409087  [14080/28709]\n",
            "loss: 0.668146  [14720/28709]\n",
            "loss: 0.694127  [15360/28709]\n",
            "loss: 0.629915  [16000/28709]\n",
            "loss: 0.645570  [16640/28709]\n",
            "loss: 0.788969  [17280/28709]\n",
            "loss: 0.674392  [17920/28709]\n",
            "loss: 0.599568  [18560/28709]\n",
            "loss: 0.604940  [19200/28709]\n",
            "loss: 0.947651  [19840/28709]\n",
            "loss: 0.683567  [20480/28709]\n",
            "loss: 0.764317  [21120/28709]\n",
            "loss: 0.589706  [21760/28709]\n",
            "loss: 0.565024  [22400/28709]\n",
            "loss: 0.568127  [23040/28709]\n",
            "loss: 0.739050  [23680/28709]\n",
            "loss: 0.868354  [24320/28709]\n",
            "loss: 0.593028  [24960/28709]\n",
            "loss: 0.513350  [25600/28709]\n",
            "loss: 0.617272  [26240/28709]\n",
            "loss: 0.854816  [26880/28709]\n",
            "loss: 0.548218  [27520/28709]\n",
            "loss: 0.638087  [28160/28709]\n",
            "\n",
            "Test Error\n",
            "\tAccuracy: 58.8%, Avg loss: 1.234462 \n",
            "\n",
            "Epoch 53\n",
            "-------------------------------\n",
            "loss: 0.588291  [    0/28709]\n",
            "loss: 0.532099  [  640/28709]\n",
            "loss: 0.600913  [ 1280/28709]\n",
            "loss: 0.686421  [ 1920/28709]\n",
            "loss: 0.547250  [ 2560/28709]\n",
            "loss: 0.764872  [ 3200/28709]\n",
            "loss: 0.636333  [ 3840/28709]\n",
            "loss: 0.774263  [ 4480/28709]\n",
            "loss: 0.659230  [ 5120/28709]\n",
            "loss: 0.663904  [ 5760/28709]\n",
            "loss: 0.674672  [ 6400/28709]\n",
            "loss: 0.524828  [ 7040/28709]\n",
            "loss: 0.629554  [ 7680/28709]\n",
            "loss: 0.674907  [ 8320/28709]\n",
            "loss: 0.771987  [ 8960/28709]\n",
            "loss: 0.758249  [ 9600/28709]\n",
            "loss: 0.589177  [10240/28709]\n",
            "loss: 0.866466  [10880/28709]\n",
            "loss: 0.523207  [11520/28709]\n",
            "loss: 0.706653  [12160/28709]\n",
            "loss: 0.508372  [12800/28709]\n",
            "loss: 0.818596  [13440/28709]\n",
            "loss: 0.878543  [14080/28709]\n",
            "loss: 0.830745  [14720/28709]\n",
            "loss: 0.691723  [15360/28709]\n",
            "loss: 0.570568  [16000/28709]\n",
            "loss: 0.591092  [16640/28709]\n",
            "loss: 0.638098  [17280/28709]\n",
            "loss: 1.005805  [17920/28709]\n",
            "loss: 0.709821  [18560/28709]\n",
            "loss: 0.727096  [19200/28709]\n",
            "loss: 0.845620  [19840/28709]\n",
            "loss: 0.758916  [20480/28709]\n",
            "loss: 0.543233  [21120/28709]\n",
            "loss: 0.700812  [21760/28709]\n",
            "loss: 0.601650  [22400/28709]\n",
            "loss: 0.729299  [23040/28709]\n",
            "loss: 0.699906  [23680/28709]\n",
            "loss: 0.840138  [24320/28709]\n",
            "loss: 0.595573  [24960/28709]\n",
            "loss: 0.480205  [25600/28709]\n",
            "loss: 0.695149  [26240/28709]\n",
            "loss: 0.573380  [26880/28709]\n",
            "loss: 0.616379  [27520/28709]\n",
            "loss: 0.616550  [28160/28709]\n",
            "\n",
            "Test Error\n",
            "\tAccuracy: 58.4%, Avg loss: 1.245394 \n",
            "\n",
            "Epoch 54\n",
            "-------------------------------\n",
            "loss: 0.445453  [    0/28709]\n",
            "loss: 0.559539  [  640/28709]\n",
            "loss: 0.577699  [ 1280/28709]\n",
            "loss: 0.565741  [ 1920/28709]\n",
            "loss: 0.634122  [ 2560/28709]\n",
            "loss: 0.786395  [ 3200/28709]\n",
            "loss: 0.556527  [ 3840/28709]\n",
            "loss: 0.545363  [ 4480/28709]\n",
            "loss: 0.717291  [ 5120/28709]\n",
            "loss: 0.599314  [ 5760/28709]\n",
            "loss: 0.561525  [ 6400/28709]\n",
            "loss: 0.631250  [ 7040/28709]\n",
            "loss: 0.562733  [ 7680/28709]\n",
            "loss: 0.499896  [ 8320/28709]\n",
            "loss: 0.789928  [ 8960/28709]\n",
            "loss: 0.608484  [ 9600/28709]\n",
            "loss: 0.706045  [10240/28709]\n",
            "loss: 0.712228  [10880/28709]\n",
            "loss: 0.646739  [11520/28709]\n",
            "loss: 0.756264  [12160/28709]\n",
            "loss: 0.756387  [12800/28709]\n",
            "loss: 0.749077  [13440/28709]\n",
            "loss: 0.745842  [14080/28709]\n",
            "loss: 0.701606  [14720/28709]\n",
            "loss: 0.809087  [15360/28709]\n",
            "loss: 0.534620  [16000/28709]\n",
            "loss: 0.761805  [16640/28709]\n",
            "loss: 0.690822  [17280/28709]\n",
            "loss: 0.475505  [17920/28709]\n",
            "loss: 0.617115  [18560/28709]\n",
            "loss: 0.615927  [19200/28709]\n",
            "loss: 0.572265  [19840/28709]\n",
            "loss: 0.676553  [20480/28709]\n",
            "loss: 0.651507  [21120/28709]\n",
            "loss: 0.785958  [21760/28709]\n",
            "loss: 0.437065  [22400/28709]\n",
            "loss: 0.604807  [23040/28709]\n",
            "loss: 0.739319  [23680/28709]\n",
            "loss: 0.559932  [24320/28709]\n",
            "loss: 0.665684  [24960/28709]\n",
            "loss: 0.870825  [25600/28709]\n",
            "loss: 0.757289  [26240/28709]\n",
            "loss: 0.600096  [26880/28709]\n",
            "loss: 0.564196  [27520/28709]\n",
            "loss: 0.667093  [28160/28709]\n",
            "\n",
            "Test Error\n",
            "\tAccuracy: 58.3%, Avg loss: 1.253605 \n",
            "\n",
            "Epoch 55\n",
            "-------------------------------\n",
            "loss: 0.623838  [    0/28709]\n",
            "loss: 0.636169  [  640/28709]\n",
            "loss: 0.573880  [ 1280/28709]\n",
            "loss: 0.729134  [ 1920/28709]\n",
            "loss: 0.718155  [ 2560/28709]\n",
            "loss: 0.627168  [ 3200/28709]\n",
            "loss: 0.574356  [ 3840/28709]\n",
            "loss: 0.722979  [ 4480/28709]\n",
            "loss: 0.585937  [ 5120/28709]\n",
            "loss: 0.767247  [ 5760/28709]\n",
            "loss: 0.783771  [ 6400/28709]\n",
            "loss: 0.671581  [ 7040/28709]\n",
            "loss: 0.820417  [ 7680/28709]\n",
            "loss: 0.526143  [ 8320/28709]\n",
            "loss: 0.644193  [ 8960/28709]\n",
            "loss: 0.503663  [ 9600/28709]\n",
            "loss: 0.727762  [10240/28709]\n",
            "loss: 0.717584  [10880/28709]\n",
            "loss: 0.788510  [11520/28709]\n",
            "loss: 0.675029  [12160/28709]\n",
            "loss: 0.558707  [12800/28709]\n",
            "loss: 0.708489  [13440/28709]\n",
            "loss: 0.763913  [14080/28709]\n",
            "loss: 0.647138  [14720/28709]\n",
            "loss: 0.809988  [15360/28709]\n",
            "loss: 0.508488  [16000/28709]\n",
            "loss: 0.560054  [16640/28709]\n",
            "loss: 0.721555  [17280/28709]\n",
            "loss: 0.762949  [17920/28709]\n",
            "loss: 0.824029  [18560/28709]\n",
            "loss: 0.563202  [19200/28709]\n",
            "loss: 0.410329  [19840/28709]\n",
            "loss: 0.782276  [20480/28709]\n",
            "loss: 0.761657  [21120/28709]\n",
            "loss: 0.556430  [21760/28709]\n",
            "loss: 0.524753  [22400/28709]\n",
            "loss: 0.616076  [23040/28709]\n",
            "loss: 0.715911  [23680/28709]\n",
            "loss: 0.693294  [24320/28709]\n",
            "loss: 0.590530  [24960/28709]\n",
            "loss: 0.736748  [25600/28709]\n",
            "loss: 0.696501  [26240/28709]\n",
            "loss: 0.627391  [26880/28709]\n",
            "loss: 0.708402  [27520/28709]\n",
            "loss: 0.594910  [28160/28709]\n",
            "\n",
            "Test Error\n",
            "\tAccuracy: 58.1%, Avg loss: 1.256082 \n",
            "\n",
            "Epoch 56\n",
            "-------------------------------\n",
            "loss: 0.537782  [    0/28709]\n",
            "loss: 0.485406  [  640/28709]\n",
            "loss: 0.594132  [ 1280/28709]\n",
            "loss: 0.549315  [ 1920/28709]\n",
            "loss: 0.558691  [ 2560/28709]\n",
            "loss: 0.707814  [ 3200/28709]\n",
            "loss: 0.492096  [ 3840/28709]\n",
            "loss: 0.841009  [ 4480/28709]\n",
            "loss: 0.723150  [ 5120/28709]\n",
            "loss: 0.506980  [ 5760/28709]\n",
            "loss: 0.573197  [ 6400/28709]\n",
            "loss: 0.739774  [ 7040/28709]\n",
            "loss: 0.661486  [ 7680/28709]\n",
            "loss: 0.515250  [ 8320/28709]\n",
            "loss: 0.622756  [ 8960/28709]\n",
            "loss: 0.526679  [ 9600/28709]\n",
            "loss: 0.567247  [10240/28709]\n",
            "loss: 0.614459  [10880/28709]\n",
            "loss: 0.471005  [11520/28709]\n",
            "loss: 0.683866  [12160/28709]\n",
            "loss: 0.837071  [12800/28709]\n",
            "loss: 0.647309  [13440/28709]\n",
            "loss: 0.709543  [14080/28709]\n",
            "loss: 0.914027  [14720/28709]\n",
            "loss: 0.823831  [15360/28709]\n",
            "loss: 0.738158  [16000/28709]\n",
            "loss: 0.573623  [16640/28709]\n",
            "loss: 0.695974  [17280/28709]\n",
            "loss: 0.716683  [17920/28709]\n",
            "loss: 0.670549  [18560/28709]\n",
            "loss: 0.713505  [19200/28709]\n",
            "loss: 0.417414  [19840/28709]\n",
            "loss: 0.585199  [20480/28709]\n",
            "loss: 0.772516  [21120/28709]\n",
            "loss: 0.650093  [21760/28709]\n",
            "loss: 0.690126  [22400/28709]\n",
            "loss: 0.448784  [23040/28709]\n",
            "loss: 0.830263  [23680/28709]\n",
            "loss: 0.547546  [24320/28709]\n",
            "loss: 0.853042  [24960/28709]\n",
            "loss: 0.643809  [25600/28709]\n",
            "loss: 0.523062  [26240/28709]\n",
            "loss: 0.605974  [26880/28709]\n",
            "loss: 0.681076  [27520/28709]\n",
            "loss: 0.557367  [28160/28709]\n",
            "\n",
            "Test Error\n",
            "\tAccuracy: 58.5%, Avg loss: 1.277980 \n",
            "\n",
            "Epoch 57\n",
            "-------------------------------\n",
            "loss: 0.530502  [    0/28709]\n",
            "loss: 0.619713  [  640/28709]\n",
            "loss: 0.542278  [ 1280/28709]\n",
            "loss: 0.540053  [ 1920/28709]\n",
            "loss: 0.611737  [ 2560/28709]\n",
            "loss: 0.512723  [ 3200/28709]\n",
            "loss: 0.548441  [ 3840/28709]\n",
            "loss: 0.687935  [ 4480/28709]\n",
            "loss: 0.441968  [ 5120/28709]\n",
            "loss: 0.473639  [ 5760/28709]\n",
            "loss: 0.689865  [ 6400/28709]\n",
            "loss: 0.581139  [ 7040/28709]\n",
            "loss: 0.717771  [ 7680/28709]\n",
            "loss: 0.765315  [ 8320/28709]\n",
            "loss: 0.655071  [ 8960/28709]\n",
            "loss: 0.827485  [ 9600/28709]\n",
            "loss: 0.601166  [10240/28709]\n",
            "loss: 0.491004  [10880/28709]\n",
            "loss: 0.636886  [11520/28709]\n",
            "loss: 0.518216  [12160/28709]\n",
            "loss: 0.424340  [12800/28709]\n",
            "loss: 0.521710  [13440/28709]\n",
            "loss: 0.727574  [14080/28709]\n",
            "loss: 0.654132  [14720/28709]\n",
            "loss: 0.725079  [15360/28709]\n",
            "loss: 0.616335  [16000/28709]\n",
            "loss: 0.896724  [16640/28709]\n",
            "loss: 0.451876  [17280/28709]\n",
            "loss: 0.563958  [17920/28709]\n",
            "loss: 0.631156  [18560/28709]\n",
            "loss: 0.768215  [19200/28709]\n",
            "loss: 0.484337  [19840/28709]\n",
            "loss: 0.763810  [20480/28709]\n",
            "loss: 0.803271  [21120/28709]\n",
            "loss: 0.661313  [21760/28709]\n",
            "loss: 0.862446  [22400/28709]\n",
            "loss: 0.751737  [23040/28709]\n",
            "loss: 0.843021  [23680/28709]\n",
            "loss: 0.634254  [24320/28709]\n",
            "loss: 0.538741  [24960/28709]\n",
            "loss: 0.588898  [25600/28709]\n",
            "loss: 0.508057  [26240/28709]\n",
            "loss: 0.761509  [26880/28709]\n",
            "loss: 0.532962  [27520/28709]\n",
            "loss: 0.568640  [28160/28709]\n",
            "\n",
            "Test Error\n",
            "\tAccuracy: 58.9%, Avg loss: 1.282081 \n",
            "\n",
            "Epoch 58\n",
            "-------------------------------\n",
            "loss: 0.709144  [    0/28709]\n",
            "loss: 0.709424  [  640/28709]\n",
            "loss: 0.620862  [ 1280/28709]\n",
            "loss: 0.601164  [ 1920/28709]\n",
            "loss: 0.454599  [ 2560/28709]\n",
            "loss: 0.876507  [ 3200/28709]\n",
            "loss: 0.598109  [ 3840/28709]\n",
            "loss: 0.578742  [ 4480/28709]\n",
            "loss: 0.464965  [ 5120/28709]\n",
            "loss: 0.458717  [ 5760/28709]\n",
            "loss: 0.920100  [ 6400/28709]\n",
            "loss: 1.063196  [ 7040/28709]\n",
            "loss: 0.513672  [ 7680/28709]\n",
            "loss: 0.712811  [ 8320/28709]\n",
            "loss: 0.627574  [ 8960/28709]\n",
            "loss: 0.716316  [ 9600/28709]\n",
            "loss: 0.469251  [10240/28709]\n",
            "loss: 0.583303  [10880/28709]\n",
            "loss: 0.815356  [11520/28709]\n",
            "loss: 0.711694  [12160/28709]\n",
            "loss: 0.627098  [12800/28709]\n",
            "loss: 0.504712  [13440/28709]\n",
            "loss: 0.646661  [14080/28709]\n",
            "loss: 0.774867  [14720/28709]\n",
            "loss: 0.506002  [15360/28709]\n",
            "loss: 0.441961  [16000/28709]\n",
            "loss: 0.621307  [16640/28709]\n",
            "loss: 0.436840  [17280/28709]\n",
            "loss: 0.514861  [17920/28709]\n",
            "loss: 0.656354  [18560/28709]\n",
            "loss: 0.632030  [19200/28709]\n",
            "loss: 0.550703  [19840/28709]\n",
            "loss: 0.575605  [20480/28709]\n",
            "loss: 0.647860  [21120/28709]\n",
            "loss: 0.874810  [21760/28709]\n",
            "loss: 0.453515  [22400/28709]\n",
            "loss: 0.724370  [23040/28709]\n",
            "loss: 0.759114  [23680/28709]\n",
            "loss: 0.453465  [24320/28709]\n",
            "loss: 0.487872  [24960/28709]\n",
            "loss: 0.666485  [25600/28709]\n",
            "loss: 0.711317  [26240/28709]\n",
            "loss: 0.545399  [26880/28709]\n",
            "loss: 0.629115  [27520/28709]\n",
            "loss: 0.588420  [28160/28709]\n",
            "\n",
            "Test Error\n",
            "\tAccuracy: 58.4%, Avg loss: 1.271837 \n",
            "\n",
            "Epoch 59\n",
            "-------------------------------\n",
            "loss: 0.688165  [    0/28709]\n",
            "loss: 0.517524  [  640/28709]\n",
            "loss: 0.566565  [ 1280/28709]\n",
            "loss: 0.547665  [ 1920/28709]\n",
            "loss: 0.764165  [ 2560/28709]\n",
            "loss: 0.712254  [ 3200/28709]\n",
            "loss: 0.554656  [ 3840/28709]\n",
            "loss: 0.701388  [ 4480/28709]\n",
            "loss: 0.619587  [ 5120/28709]\n",
            "loss: 0.811621  [ 5760/28709]\n",
            "loss: 0.652465  [ 6400/28709]\n",
            "loss: 0.586430  [ 7040/28709]\n",
            "loss: 0.699818  [ 7680/28709]\n",
            "loss: 0.521844  [ 8320/28709]\n",
            "loss: 0.633827  [ 8960/28709]\n",
            "loss: 0.537126  [ 9600/28709]\n",
            "loss: 0.774529  [10240/28709]\n",
            "loss: 0.569453  [10880/28709]\n",
            "loss: 0.528745  [11520/28709]\n",
            "loss: 0.678713  [12160/28709]\n",
            "loss: 0.552081  [12800/28709]\n",
            "loss: 0.806756  [13440/28709]\n",
            "loss: 0.535374  [14080/28709]\n",
            "loss: 0.706630  [14720/28709]\n",
            "loss: 0.565060  [15360/28709]\n",
            "loss: 0.676247  [16000/28709]\n",
            "loss: 0.482201  [16640/28709]\n",
            "loss: 0.411226  [17280/28709]\n",
            "loss: 0.722202  [17920/28709]\n",
            "loss: 0.500120  [18560/28709]\n",
            "loss: 0.757496  [19200/28709]\n",
            "loss: 0.853922  [19840/28709]\n",
            "loss: 0.775421  [20480/28709]\n",
            "loss: 0.397817  [21120/28709]\n",
            "loss: 0.625651  [21760/28709]\n",
            "loss: 0.635015  [22400/28709]\n",
            "loss: 0.869657  [23040/28709]\n",
            "loss: 0.729947  [23680/28709]\n",
            "loss: 0.617650  [24320/28709]\n",
            "loss: 0.536422  [24960/28709]\n",
            "loss: 0.432117  [25600/28709]\n",
            "loss: 0.751981  [26240/28709]\n",
            "loss: 0.691422  [26880/28709]\n",
            "loss: 0.655444  [27520/28709]\n",
            "loss: 0.820480  [28160/28709]\n",
            "\n",
            "Test Error\n",
            "\tAccuracy: 58.3%, Avg loss: 1.279108 \n",
            "\n",
            "Epoch 60\n",
            "-------------------------------\n",
            "loss: 0.542549  [    0/28709]\n",
            "loss: 0.551846  [  640/28709]\n",
            "loss: 0.652249  [ 1280/28709]\n",
            "loss: 0.461131  [ 1920/28709]\n",
            "loss: 0.501163  [ 2560/28709]\n",
            "loss: 0.576358  [ 3200/28709]\n",
            "loss: 0.490056  [ 3840/28709]\n",
            "loss: 0.575068  [ 4480/28709]\n",
            "loss: 0.608198  [ 5120/28709]\n",
            "loss: 0.542850  [ 5760/28709]\n",
            "loss: 0.799838  [ 6400/28709]\n",
            "loss: 0.611845  [ 7040/28709]\n",
            "loss: 0.578522  [ 7680/28709]\n",
            "loss: 0.602502  [ 8320/28709]\n",
            "loss: 0.583064  [ 8960/28709]\n",
            "loss: 0.592980  [ 9600/28709]\n",
            "loss: 0.447557  [10240/28709]\n",
            "loss: 0.761440  [10880/28709]\n",
            "loss: 0.626752  [11520/28709]\n",
            "loss: 0.881606  [12160/28709]\n",
            "loss: 0.593235  [12800/28709]\n",
            "loss: 0.676891  [13440/28709]\n",
            "loss: 0.458188  [14080/28709]\n",
            "loss: 0.442199  [14720/28709]\n",
            "loss: 0.704387  [15360/28709]\n",
            "loss: 0.547042  [16000/28709]\n",
            "loss: 0.517555  [16640/28709]\n",
            "loss: 0.449450  [17280/28709]\n",
            "loss: 0.451390  [17920/28709]\n",
            "loss: 0.559984  [18560/28709]\n",
            "loss: 0.641126  [19200/28709]\n",
            "loss: 0.742009  [19840/28709]\n",
            "loss: 0.667131  [20480/28709]\n",
            "loss: 0.703659  [21120/28709]\n",
            "loss: 0.656725  [21760/28709]\n",
            "loss: 0.802360  [22400/28709]\n",
            "loss: 0.587352  [23040/28709]\n",
            "loss: 0.759652  [23680/28709]\n",
            "loss: 0.784947  [24320/28709]\n",
            "loss: 0.507717  [24960/28709]\n",
            "loss: 0.684508  [25600/28709]\n",
            "loss: 0.893628  [26240/28709]\n",
            "loss: 0.620218  [26880/28709]\n",
            "loss: 0.948985  [27520/28709]\n",
            "loss: 0.553927  [28160/28709]\n",
            "\n",
            "Test Error\n",
            "\tAccuracy: 58.3%, Avg loss: 1.260914 \n",
            "\n",
            "Saving model to disk!\n"
          ]
        }
      ],
      "source": [
        "def model_train_loader(model, dataloader, criterion, optimizer):\n",
        "    size = len(dataloader.dataset)\n",
        "\n",
        "    model.train()\n",
        "    for batch, (X, y) in enumerate(dataloader):\n",
        "        # Compute prediction error\n",
        "        X=X.to(device)\n",
        "        y=y.to(device)\n",
        "        predictions = model(X)\n",
        "        loss = criterion(predictions, y)\n",
        "\n",
        "        # Backpropagation\n",
        "        optimizer.zero_grad()\n",
        "        loss.backward()\n",
        "        optimizer.step()\n",
        "\n",
        "        # Report loss every 10 batches\n",
        "        if batch % 10 == 0:\n",
        "            loss, current = loss.item(), batch * len(X)\n",
        "            print(f\"loss: {loss:>7f}  [{current:>5d}/{size:>5d}]\")\n",
        "    \n",
        "            \n",
        "def model_test_loader(model, dataloader, criterion):\n",
        "    size = len(dataloader.dataset)\n",
        "    num_batches = len(dataloader)\n",
        "    # Tracking test loss (cross-entropy) and correct classification rate (accuracy)\n",
        "    test_loss, correct = 0, 0\n",
        "    \n",
        "    with torch.no_grad():\n",
        "        model.eval()\n",
        "        for X, y in dataloader:\n",
        "            X=X.to(device)\n",
        "            y=y.to(device)\n",
        "            predictions = model(X)\n",
        "            test_loss += criterion(predictions, y)\n",
        "            correct += (predictions.argmax(1) == y).type(torch.float).sum().item()\n",
        "            \n",
        "    test_loss /= num_batches\n",
        "    correct /= size\n",
        "    print(f\"\\nTest Error\\n\\tAccuracy: {(100*correct):>0.1f}%, Avg loss: {test_loss:>8f} \\n\")\n",
        "\n",
        "# The nn.CrossEntropyLoss() loss function automatically performs a log_softmax() to output\n",
        "criterion = nn.CrossEntropyLoss()\n",
        "# One is enough for my toy example...\n",
        "num_epochs = 60\n",
        "optimizer = torch.optim.Adam(model.parameters(), lr=0.00008)#, momentum=0.9, nesterov=True)\n",
        "\n",
        "for t in range(num_epochs):\n",
        "    print(f\"Epoch {t+1}\\n-------------------------------\")\n",
        "    model_train_loader(model, train_dataloader, criterion, optimizer)\n",
        "    model_test_loader(model, test_dataloader, criterion)\n",
        "    \n",
        "# Saving the model file 'cnn_fmd.pth' to my current working directory (cwd)\n",
        "print(\"Saving model to disk!\")\n",
        "torch.save(model.state_dict(), os.getcwd() + '/cnn_fmd.pth')"
      ],
      "id": "DTe-UYFiV2vo"
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "96dc7980",
        "outputId": "69533a4e-943e-42f1-b777-54331df521f9"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "\n",
            "GroundTruth:\n",
            " neutral\t fear \t sad  \t happy\t happy\t surprise\t happy\t neutral\t\n",
            "\n",
            "Predicted:\n",
            " neutral\t surprise\t neutral\t happy\t fear \t happy\t happy\t happy\t\n"
          ]
        },
        {
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 1440x1440 with 1 Axes>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        }
      ],
      "source": [
        "### Plotting stuff ###\n",
        "fig = plt.figure(figsize=(20,20))\n",
        "\n",
        "# Get some random test dataset images\n",
        "dataiter = iter(test_dataloader)\n",
        "# Extract a batch\n",
        "images, labels = dataiter.next()\n",
        "# Show images\n",
        "imshow(torchvision.utils.make_grid(images[:num_display]))\n",
        "\n",
        "outputs = model(images)\n",
        "# Argmax take the most probable class\n",
        "_, predicted = torch.max(outputs, 1)\n",
        "\n",
        "# Print labels\n",
        "print(\"\\nGroundTruth:\\n\", \" \".join(f'{classes[labels[j]]:5s}\\t' for j in range(8)))\n",
        "\n",
        "print(\"\\nPredicted:\\n\", \" \".join(f'{classes[predicted[j]]:5s}\\t' for j in range(8)))"
      ],
      "id": "96dc7980"
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "d502a756",
        "outputId": "7e3a7224-5611-46d4-952a-0f280d2283c9"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Accuracy of the network on the test set images: 37 %\n",
            "Confusion Matrix (columns: True class, rows: Predicted class):\n"
          ]
        },
        {
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        }
      ],
      "source": [
        "correct = 0\n",
        "\n",
        "predicted_list = []\n",
        "labels_list = []\n",
        "\n",
        "# Since we're not training, we don't need to calculate the gradients for our outputs\n",
        "with torch.no_grad():\n",
        "    i = 0\n",
        "    for data in test_dataloader:\n",
        "        i += 1\n",
        "        images, labels = data\n",
        "        # Forward pass through network\n",
        "        outputs = model(images)\n",
        "        # Take most probable class\n",
        "        _, predicted = torch.max(outputs.data, 1)\n",
        "        predicted_list.append(predicted.detach().numpy())\n",
        "        labels_list.append(labels.detach().numpy())\n",
        "        \n",
        "# Hacky conversions from a list of lists of arrays into joined numpy arrays\n",
        "predicted_list = np.array(list(itertools.chain.from_iterable(predicted_list)))\n",
        "labels_list = np.array(list(itertools.chain.from_iterable(labels_list)))\n",
        "# Count up number of labels and correct predictions\n",
        "correct = sum(predicted_list == labels_list)\n",
        "\n",
        "print(f\"Accuracy of the network on the test set images: {100 * correct // len(test_dataloader.dataset)} %\")\n",
        "\n",
        "print(\"Confusion Matrix (columns: True class, rows: Predicted class):\")\n",
        "conf_mat = confusion_matrix(labels_list, predicted_list);\n",
        "conf_display = ConfusionMatrixDisplay.from_predictions(labels_list, predicted_list, display_labels=classes, colorbar=False);\n",
        "plt.xlabel(\"True Labels\");\n",
        "plt.ylabel(\"Predicted Labels\");"
      ],
      "id": "d502a756"
    }
  ],
  "metadata": {
    "accelerator": "GPU",
    "colab": {
      "collapsed_sections": [],
      "machine_shape": "hm",
      "name": "Final Project.ipynb",
      "provenance": []
    },
    "gpuClass": "standard",
    "kernelspec": {
      "display_name": "Python 3 (ipykernel)",
      "language": "python",
      "name": "python3"
    },
    "language_info": {
      "codemirror_mode": {
        "name": "ipython",
        "version": 3
      },
      "file_extension": ".py",
      "mimetype": "text/x-python",
      "name": "python",
      "nbconvert_exporter": "python",
      "pygments_lexer": "ipython3",
      "version": "3.9.7"
    }
  },
  "nbformat": 4,
  "nbformat_minor": 5
}