DL_Project_overfit.ipynb

{
  "cells": [
    {
      "cell_type": "code",
      "execution_count": 1,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "W9G6tHkfFcT_",
        "outputId": "3cd30e13-4795-484d-8d70-44750487bf55"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Mounted at /content/drive\n"
          ]
        }
      ],
      "source": [
        "from google.colab import drive\n",
        "drive.mount('/content/drive')"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 2,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "wcwF5uw-r1UY",
        "outputId": "c9102e73-a9dc-47ff-ce5e-26a3626308e4"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Archive:  /content/drive/MyDrive/minist_overfit.zip\n",
            "   creating: data/test/\n",
            "   creating: data/test/0/\n",
            " extracting: data/test/0/9802.png    \n",
            " extracting: data/test/0/9810.png    \n",
            " extracting: data/test/0/9819.png    \n",
            " extracting: data/test/0/9827.png    \n",
            " extracting: data/test/0/9851.png    \n",
            " extracting: data/test/0/9866.png    \n",
            " extracting: data/test/0/9880.png    \n",
            " extracting: data/test/0/9886.png    \n",
            " extracting: data/test/0/9898.png    \n",
            " extracting: data/test/0/9903.png    \n",
            " extracting: data/test/0/9912.png    \n",
            " extracting: data/test/0/9931.png    \n",
            " extracting: data/test/0/9940.png    \n",
            " extracting: data/test/0/9949.png    \n",
            "  inflating: data/test/0/9952.png    \n",
            " extracting: data/test/0/9953.png    \n",
            " extracting: data/test/0/9963.png    \n",
            " extracting: data/test/0/9965.png    \n",
            " extracting: data/test/0/9984.png    \n",
            " extracting: data/test/0/9994.png    \n",
            "   creating: data/test/1/\n",
            " extracting: data/test/1/7807.png    \n",
            " extracting: data/test/1/7812.png    \n",
            " extracting: data/test/1/7813.png    \n",
            " extracting: data/test/1/7823.png    \n",
            " extracting: data/test/1/7833.png    \n",
            " extracting: data/test/1/7840.png    \n",
            " extracting: data/test/1/7848.png    \n",
            " extracting: data/test/1/7857.png    \n",
            " extracting: data/test/1/7870.png    \n",
            " extracting: data/test/1/7882.png    \n",
            " extracting: data/test/1/7891.png    \n",
            " extracting: data/test/1/7897.png    \n",
            " extracting: data/test/1/7900.png    \n",
            " extracting: data/test/1/7901.png    \n",
            " extracting: data/test/1/7912.png    \n",
            " extracting: data/test/1/7917.png    \n",
            " extracting: data/test/1/7921.png    \n",
            " extracting: data/test/1/7929.png    \n",
            "  inflating: data/test/1/7935.png    \n",
            "  inflating: data/test/1/7945.png    \n",
            "   creating: data/test/2/\n",
            " extracting: data/test/2/1057.png    \n",
            " extracting: data/test/2/1178.png    \n",
            "  inflating: data/test/2/1188.png    \n",
            " extracting: data/test/2/150.png     \n",
            "  inflating: data/test/2/304.png     \n",
            " extracting: data/test/2/401.png     \n",
            " extracting: data/test/2/445.png     \n",
            " extracting: data/test/2/514.png     \n",
            " extracting: data/test/2/540.png     \n",
            " extracting: data/test/2/630.png     \n",
            " extracting: data/test/2/633.png     \n",
            " extracting: data/test/2/636.png     \n",
            " extracting: data/test/2/647.png     \n",
            "  inflating: data/test/2/660.png     \n",
            " extracting: data/test/2/864.png     \n",
            " extracting: data/test/2/876.png     \n",
            " extracting: data/test/2/923.png     \n",
            " extracting: data/test/2/927.png     \n",
            " extracting: data/test/2/975.png     \n",
            " extracting: data/test/2/996.png     \n",
            "   creating: data/test/3/\n",
            " extracting: data/test/3/174.png     \n",
            "  inflating: data/test/3/196.png     \n",
            "  inflating: data/test/3/201.png     \n",
            " extracting: data/test/3/246.png     \n",
            " extracting: data/test/3/2534.png    \n",
            " extracting: data/test/3/2619.png    \n",
            " extracting: data/test/3/2630.png    \n",
            " extracting: data/test/3/5447.png    \n",
            " extracting: data/test/3/5688.png    \n",
            " extracting: data/test/3/5701.png    \n",
            " extracting: data/test/3/7208.png    \n",
            " extracting: data/test/3/7212.png    \n",
            " extracting: data/test/3/7234.png    \n",
            " extracting: data/test/3/7258.png    \n",
            " extracting: data/test/3/7596.png    \n",
            " extracting: data/test/3/7842.png    \n",
            " extracting: data/test/3/7850.png    \n",
            " extracting: data/test/3/8509.png    \n",
            " extracting: data/test/3/88.png      \n",
            " extracting: data/test/3/9343.png    \n",
            " extracting: data/test/3/9371.png    \n",
            " extracting: data/test/3/9430.png    \n",
            " extracting: data/test/3/9431.png    \n",
            " extracting: data/test/3/9480.png    \n",
            " extracting: data/test/3/9637.png    \n",
            " extracting: data/test/3/9656.png    \n",
            "   creating: data/test/4/\n",
            " extracting: data/test/4/122.png     \n",
            "  inflating: data/test/4/1438.png    \n",
            "  inflating: data/test/4/1441.png    \n",
            "  inflating: data/test/4/161.png     \n",
            "  inflating: data/test/4/2006.png    \n",
            " extracting: data/test/4/2054.png    \n",
            " extracting: data/test/4/2149.png    \n",
            " extracting: data/test/4/2339.png    \n",
            " extracting: data/test/4/2660.png    \n",
            " extracting: data/test/4/309.png     \n",
            " extracting: data/test/4/3342.png    \n",
            " extracting: data/test/4/3406.png    \n",
            " extracting: data/test/4/4195.png    \n",
            "  inflating: data/test/4/4383.png    \n",
            " extracting: data/test/4/5261.png    \n",
            "  inflating: data/test/4/5336.png    \n",
            " extracting: data/test/4/6560.png    \n",
            " extracting: data/test/4/6770.png    \n",
            "  inflating: data/test/4/8590.png    \n",
            " extracting: data/test/4/86.png      \n",
            "  inflating: data/test/4/8663.png    \n",
            " extracting: data/test/4/8887.png    \n",
            " extracting: data/test/4/96.png      \n",
            "   creating: data/test/5/\n",
            " extracting: data/test/5/16.png      \n",
            " extracting: data/test/5/168.png     \n",
            " extracting: data/test/5/241.png     \n",
            "  inflating: data/test/5/254.png     \n",
            " extracting: data/test/5/262.png     \n",
            " extracting: data/test/5/318.png     \n",
            " extracting: data/test/5/3394.png    \n",
            " extracting: data/test/5/3415.png    \n",
            " extracting: data/test/5/3624.png    \n",
            " extracting: data/test/5/3664.png    \n",
            " extracting: data/test/5/3703.png    \n",
            " extracting: data/test/5/60.png      \n",
            " extracting: data/test/5/6088.png    \n",
            " extracting: data/test/5/6156.png    \n",
            "  inflating: data/test/5/6166.png    \n",
            " extracting: data/test/5/6245.png    \n",
            " extracting: data/test/5/8854.png    \n",
            " extracting: data/test/5/8864.png    \n",
            " extracting: data/test/5/9.png       \n",
            " extracting: data/test/5/9546.png    \n",
            " extracting: data/test/5/9607.png    \n",
            " extracting: data/test/5/9750.png    \n",
            " extracting: data/test/5/9755.png    \n",
            " extracting: data/test/5/9771.png    \n",
            " extracting: data/test/5/9983.png    \n",
            "   creating: data/test/6/\n",
            " extracting: data/test/6/124.png     \n",
            " extracting: data/test/6/1308.png    \n",
            " extracting: data/test/6/1345.png    \n",
            " extracting: data/test/6/1363.png    \n",
            " extracting: data/test/6/1378.png    \n",
            " extracting: data/test/6/139.png     \n",
            " extracting: data/test/6/165.png     \n",
            " extracting: data/test/6/22.png      \n",
            " extracting: data/test/6/3068.png    \n",
            " extracting: data/test/6/3089.png    \n",
            " extracting: data/test/6/3303.png    \n",
            " extracting: data/test/6/7861.png    \n",
            " extracting: data/test/6/7898.png    \n",
            " extracting: data/test/6/7926.png    \n",
            " extracting: data/test/6/7930.png    \n",
            " extracting: data/test/6/8064.png    \n",
            " extracting: data/test/6/9022.png    \n",
            " extracting: data/test/6/9128.png    \n",
            " extracting: data/test/6/9533.png    \n",
            " extracting: data/test/6/9618.png    \n",
            " extracting: data/test/6/9659.png    \n",
            " extracting: data/test/6/9863.png    \n",
            " extracting: data/test/6/9889.png    \n",
            " extracting: data/test/6/9922.png    \n",
            "   creating: data/test/7/\n",
            " extracting: data/test/7/1123.png    \n",
            " extracting: data/test/7/1174.png    \n",
            " extracting: data/test/7/1187.png    \n",
            " extracting: data/test/7/1242.png    \n",
            " extracting: data/test/7/1245.png    \n",
            " extracting: data/test/7/1261.png    \n",
            " extracting: data/test/7/1329.png    \n",
            "  inflating: data/test/7/1348.png    \n",
            " extracting: data/test/7/1356.png    \n",
            " extracting: data/test/7/1717.png    \n",
            " extracting: data/test/7/1719.png    \n",
            " extracting: data/test/7/1784.png    \n",
            " extracting: data/test/7/1800.png    \n",
            " extracting: data/test/7/1806.png    \n",
            " extracting: data/test/7/1810.png    \n",
            "  inflating: data/test/7/2025.png    \n",
            "  inflating: data/test/7/221.png     \n",
            " extracting: data/test/7/27.png      \n",
            " extracting: data/test/7/359.png     \n",
            " extracting: data/test/7/606.png     \n",
            " extracting: data/test/7/685.png     \n",
            " extracting: data/test/7/727.png     \n",
            " extracting: data/test/7/842.png     \n",
            " extracting: data/test/7/847.png     \n",
            " extracting: data/test/7/881.png     \n",
            " extracting: data/test/7/899.png     \n",
            " extracting: data/test/7/98.png      \n",
            "   creating: data/test/8/\n",
            " extracting: data/test/8/1034.png    \n",
            " extracting: data/test/8/1069.png    \n",
            " extracting: data/test/8/1126.png    \n",
            " extracting: data/test/8/1171.png    \n",
            " extracting: data/test/8/1186.png    \n",
            " extracting: data/test/8/1365.png    \n",
            " extracting: data/test/8/1426.png    \n",
            " extracting: data/test/8/1442.png    \n",
            "  inflating: data/test/8/1531.png    \n",
            " extracting: data/test/8/1563.png    \n",
            " extracting: data/test/8/1568.png    \n",
            "  inflating: data/test/8/1585.png    \n",
            " extracting: data/test/8/2897.png    \n",
            "  inflating: data/test/8/2961.png    \n",
            " extracting: data/test/8/3024.png    \n",
            "  inflating: data/test/8/3568.png    \n",
            " extracting: data/test/8/3675.png    \n",
            " extracting: data/test/8/3758.png    \n",
            " extracting: data/test/8/4725.png    \n",
            " extracting: data/test/8/4732.png    \n",
            " extracting: data/test/8/5651.png    \n",
            " extracting: data/test/8/5856.png    \n",
            " extracting: data/test/8/9597.png    \n",
            " extracting: data/test/8/9733.png    \n",
            " extracting: data/test/8/9745.png    \n",
            " extracting: data/test/8/9888.png    \n",
            " extracting: data/test/8/9918.png    \n",
            "   creating: data/test/9/\n",
            " extracting: data/test/9/1001.png    \n",
            " extracting: data/test/9/1064.png    \n",
            " extracting: data/test/9/1108.png    \n",
            " extracting: data/test/9/1248.png    \n",
            " extracting: data/test/9/152.png     \n",
            " extracting: data/test/9/1560.png    \n",
            " extracting: data/test/9/1669.png    \n",
            "  inflating: data/test/9/1710.png    \n",
            " extracting: data/test/9/2010.png    \n",
            " extracting: data/test/9/2148.png    \n",
            " extracting: data/test/9/2162.png    \n",
            " extracting: data/test/9/2293.png    \n",
            " extracting: data/test/9/2294.png    \n",
            " extracting: data/test/9/265.png     \n",
            "  inflating: data/test/9/2736.png    \n",
            " extracting: data/test/9/2741.png    \n",
            " extracting: data/test/9/2846.png    \n",
            " extracting: data/test/9/2940.png    \n",
            " extracting: data/test/9/3002.png    \n",
            "  inflating: data/test/9/3968.png    \n",
            " extracting: data/test/9/4016.png    \n",
            "  inflating: data/test/9/4155.png    \n",
            " extracting: data/test/9/5211.png    \n",
            " extracting: data/test/9/5263.png    \n",
            " extracting: data/test/9/5270.png    \n",
            " extracting: data/test/9/5504.png    \n",
            " extracting: data/test/9/8.png       \n",
            " extracting: data/test/9/93.png      \n",
            "   creating: data/train/\n",
            "   creating: data/train/ID0/\n",
            "   creating: data/train/ID0/0/\n",
            " extracting: data/train/ID0/0/15521.png  \n",
            " extracting: data/train/ID0/0/27895.png  \n",
            " extracting: data/train/ID0/0/29337.png  \n",
            " extracting: data/train/ID0/0/29457.png  \n",
            " extracting: data/train/ID0/0/38156.png  \n",
            " extracting: data/train/ID0/0/46115.png  \n",
            " extracting: data/train/ID0/0/46140.png  \n",
            " extracting: data/train/ID0/0/4750.png  \n",
            " extracting: data/train/ID0/0/6685.png  \n",
            " extracting: data/train/ID0/0/9553.png  \n",
            "   creating: data/train/ID0/1/\n",
            "   creating: data/train/ID0/2/\n",
            " extracting: data/train/ID0/2/10074.png  \n",
            " extracting: data/train/ID0/2/24230.png  \n",
            " extracting: data/train/ID0/2/25089.png  \n",
            " extracting: data/train/ID0/2/26084.png  \n",
            " extracting: data/train/ID0/2/34309.png  \n",
            "  inflating: data/train/ID0/2/4158.png  \n",
            "  inflating: data/train/ID0/2/51531.png  \n",
            " extracting: data/train/ID0/2/54755.png  \n",
            " extracting: data/train/ID0/2/55936.png  \n",
            " extracting: data/train/ID0/2/7940.png  \n",
            "   creating: data/train/ID0/3/\n",
            " extracting: data/train/ID0/3/12081.png  \n",
            " extracting: data/train/ID0/3/12991.png  \n",
            " extracting: data/train/ID0/3/13907.png  \n",
            " extracting: data/train/ID0/3/21371.png  \n",
            " extracting: data/train/ID0/3/30242.png  \n",
            " extracting: data/train/ID0/3/33030.png  \n",
            " extracting: data/train/ID0/3/3551.png  \n",
            " extracting: data/train/ID0/3/47726.png  \n",
            " extracting: data/train/ID0/3/547.png  \n",
            " extracting: data/train/ID0/3/57915.png  \n",
            "   creating: data/train/ID0/4/\n",
            " extracting: data/train/ID0/4/22099.png  \n",
            " extracting: data/train/ID0/4/23975.png  \n",
            " extracting: data/train/ID0/4/24193.png  \n",
            " extracting: data/train/ID0/4/30530.png  \n",
            "  inflating: data/train/ID0/4/36635.png  \n",
            " extracting: data/train/ID0/4/39246.png  \n",
            " extracting: data/train/ID0/4/47938.png  \n",
            "  inflating: data/train/ID0/4/51383.png  \n",
            " extracting: data/train/ID0/4/56174.png  \n",
            " extracting: data/train/ID0/4/9627.png  \n",
            "   creating: data/train/ID0/5/\n",
            " extracting: data/train/ID0/5/19107.png  \n",
            " extracting: data/train/ID0/5/19534.png  \n",
            " extracting: data/train/ID0/5/23408.png  \n",
            " extracting: data/train/ID0/5/29617.png  \n",
            " extracting: data/train/ID0/5/40269.png  \n",
            " extracting: data/train/ID0/5/41981.png  \n",
            "  inflating: data/train/ID0/5/43022.png  \n",
            " extracting: data/train/ID0/5/52123.png  \n",
            " extracting: data/train/ID0/5/55397.png  \n",
            " extracting: data/train/ID0/5/6530.png  \n",
            "   creating: data/train/ID0/6/\n",
            " extracting: data/train/ID0/6/14661.png  \n",
            " extracting: data/train/ID0/6/16823.png  \n",
            " extracting: data/train/ID0/6/31256.png  \n",
            " extracting: data/train/ID0/6/32695.png  \n",
            " extracting: data/train/ID0/6/50083.png  \n",
            " extracting: data/train/ID0/6/50281.png  \n",
            " extracting: data/train/ID0/6/50818.png  \n",
            " extracting: data/train/ID0/6/52582.png  \n",
            " extracting: data/train/ID0/6/53216.png  \n",
            " extracting: data/train/ID0/6/57843.png  \n",
            "   creating: data/train/ID0/7/\n",
            "  inflating: data/train/ID0/7/14751.png  \n",
            " extracting: data/train/ID0/7/16595.png  \n",
            " extracting: data/train/ID0/7/23353.png  \n",
            " extracting: data/train/ID0/7/24814.png  \n",
            " extracting: data/train/ID0/7/39039.png  \n",
            "  inflating: data/train/ID0/7/45595.png  \n",
            " extracting: data/train/ID0/7/46502.png  \n",
            "  inflating: data/train/ID0/7/50731.png  \n",
            " extracting: data/train/ID0/7/50873.png  \n",
            " extracting: data/train/ID0/7/58394.png  \n",
            "   creating: data/train/ID0/8/\n",
            " extracting: data/train/ID0/8/19254.png  \n",
            " extracting: data/train/ID0/8/22552.png  \n",
            " extracting: data/train/ID0/8/25229.png  \n",
            " extracting: data/train/ID0/8/3481.png  \n",
            " extracting: data/train/ID0/8/37126.png  \n",
            " extracting: data/train/ID0/8/37934.png  \n",
            "  inflating: data/train/ID0/8/38352.png  \n",
            " extracting: data/train/ID0/8/39383.png  \n",
            " extracting: data/train/ID0/8/58179.png  \n",
            " extracting: data/train/ID0/8/59295.png  \n",
            "   creating: data/train/ID0/9/\n",
            " extracting: data/train/ID0/9/14978.png  \n",
            " extracting: data/train/ID0/9/18778.png  \n",
            " extracting: data/train/ID0/9/26109.png  \n",
            " extracting: data/train/ID0/9/2798.png  \n",
            "  inflating: data/train/ID0/9/29823.png  \n",
            " extracting: data/train/ID0/9/41876.png  \n",
            " extracting: data/train/ID0/9/42722.png  \n",
            " extracting: data/train/ID0/9/43283.png  \n",
            " extracting: data/train/ID0/9/44932.png  \n",
            " extracting: data/train/ID0/9/58617.png  \n",
            "   creating: data/train/ID1/\n",
            "   creating: data/train/ID1/0/\n",
            " extracting: data/train/ID1/0/11792.png  \n",
            " extracting: data/train/ID1/0/14020.png  \n",
            " extracting: data/train/ID1/0/17933.png  \n",
            " extracting: data/train/ID1/0/22024.png  \n",
            " extracting: data/train/ID1/0/36895.png  \n",
            " extracting: data/train/ID1/0/48188.png  \n",
            " extracting: data/train/ID1/0/52227.png  \n",
            " extracting: data/train/ID1/0/53536.png  \n",
            " extracting: data/train/ID1/0/56354.png  \n",
            " extracting: data/train/ID1/0/850.png  \n",
            "   creating: data/train/ID1/1/\n",
            " extracting: data/train/ID1/1/1788.png  \n",
            "  inflating: data/train/ID1/1/18283.png  \n",
            "  inflating: data/train/ID1/1/24607.png  \n",
            "  inflating: data/train/ID1/1/2655.png  \n",
            " extracting: data/train/ID1/1/31740.png  \n",
            "  inflating: data/train/ID1/1/43972.png  \n",
            " extracting: data/train/ID1/1/52382.png  \n",
            "  inflating: data/train/ID1/1/57517.png  \n",
            " extracting: data/train/ID1/1/57715.png  \n",
            " extracting: data/train/ID1/1/59693.png  \n",
            "   creating: data/train/ID1/2/\n",
            "   creating: data/train/ID1/3/\n",
            " extracting: data/train/ID1/3/17083.png  \n",
            " extracting: data/train/ID1/3/1779.png  \n",
            " extracting: data/train/ID1/3/21994.png  \n",
            " extracting: data/train/ID1/3/29900.png  \n",
            " extracting: data/train/ID1/3/31033.png  \n",
            " extracting: data/train/ID1/3/31806.png  \n",
            " extracting: data/train/ID1/3/32795.png  \n",
            " extracting: data/train/ID1/3/45006.png  \n",
            " extracting: data/train/ID1/3/46280.png  \n",
            " extracting: data/train/ID1/3/56438.png  \n",
            "   creating: data/train/ID1/4/\n",
            " extracting: data/train/ID1/4/10463.png  \n",
            "  inflating: data/train/ID1/4/18366.png  \n",
            "  inflating: data/train/ID1/4/27801.png  \n",
            " extracting: data/train/ID1/4/37737.png  \n",
            "  inflating: data/train/ID1/4/40427.png  \n",
            " extracting: data/train/ID1/4/47844.png  \n",
            " extracting: data/train/ID1/4/51126.png  \n",
            " extracting: data/train/ID1/4/51262.png  \n",
            " extracting: data/train/ID1/4/648.png  \n",
            " extracting: data/train/ID1/4/693.png  \n",
            "   creating: data/train/ID1/5/\n",
            " extracting: data/train/ID1/5/10278.png  \n",
            " extracting: data/train/ID1/5/16746.png  \n",
            "  inflating: data/train/ID1/5/26843.png  \n",
            " extracting: data/train/ID1/5/28260.png  \n",
            "  inflating: data/train/ID1/5/35042.png  \n",
            " extracting: data/train/ID1/5/38935.png  \n",
            "  inflating: data/train/ID1/5/39965.png  \n",
            " extracting: data/train/ID1/5/51737.png  \n",
            " extracting: data/train/ID1/5/53696.png  \n",
            " extracting: data/train/ID1/5/53972.png  \n",
            "   creating: data/train/ID1/6/\n",
            " extracting: data/train/ID1/6/10951.png  \n",
            " extracting: data/train/ID1/6/18024.png  \n",
            " extracting: data/train/ID1/6/28808.png  \n",
            " extracting: data/train/ID1/6/32840.png  \n",
            " extracting: data/train/ID1/6/38783.png  \n",
            " extracting: data/train/ID1/6/38876.png  \n",
            " extracting: data/train/ID1/6/39825.png  \n",
            " extracting: data/train/ID1/6/43282.png  \n",
            " extracting: data/train/ID1/6/5577.png  \n",
            " extracting: data/train/ID1/6/6691.png  \n",
            "   creating: data/train/ID1/7/\n",
            " extracting: data/train/ID1/7/12600.png  \n",
            " extracting: data/train/ID1/7/13514.png  \n",
            " extracting: data/train/ID1/7/20972.png  \n",
            "  inflating: data/train/ID1/7/23591.png  \n",
            " extracting: data/train/ID1/7/25309.png  \n",
            " extracting: data/train/ID1/7/25999.png  \n",
            " extracting: data/train/ID1/7/46951.png  \n",
            " extracting: data/train/ID1/7/50580.png  \n",
            " extracting: data/train/ID1/7/53004.png  \n",
            " extracting: data/train/ID1/7/59425.png  \n",
            "   creating: data/train/ID1/8/\n",
            " extracting: data/train/ID1/8/20946.png  \n",
            " extracting: data/train/ID1/8/23738.png  \n",
            " extracting: data/train/ID1/8/26617.png  \n",
            " extracting: data/train/ID1/8/33448.png  \n",
            "  inflating: data/train/ID1/8/33515.png  \n",
            " extracting: data/train/ID1/8/38421.png  \n",
            " extracting: data/train/ID1/8/39239.png  \n",
            " extracting: data/train/ID1/8/42901.png  \n",
            " extracting: data/train/ID1/8/49053.png  \n",
            " extracting: data/train/ID1/8/49401.png  \n",
            "   creating: data/train/ID1/9/\n",
            " extracting: data/train/ID1/9/18449.png  \n",
            " extracting: data/train/ID1/9/2034.png  \n",
            " extracting: data/train/ID1/9/22407.png  \n",
            "  inflating: data/train/ID1/9/27735.png  \n",
            " extracting: data/train/ID1/9/28193.png  \n",
            "  inflating: data/train/ID1/9/45184.png  \n",
            " extracting: data/train/ID1/9/55098.png  \n",
            " extracting: data/train/ID1/9/58423.png  \n",
            " extracting: data/train/ID1/9/59206.png  \n",
            " extracting: data/train/ID1/9/9071.png  \n",
            "   creating: data/train/ID2/\n",
            "   creating: data/train/ID2/0/\n",
            " extracting: data/train/ID2/0/12671.png  \n",
            "  inflating: data/train/ID2/0/18274.png  \n",
            " extracting: data/train/ID2/0/22289.png  \n",
            " extracting: data/train/ID2/0/36810.png  \n",
            " extracting: data/train/ID2/0/46823.png  \n",
            " extracting: data/train/ID2/0/5394.png  \n",
            " extracting: data/train/ID2/0/56534.png  \n",
            " extracting: data/train/ID2/0/56804.png  \n",
            " extracting: data/train/ID2/0/59100.png  \n",
            " extracting: data/train/ID2/0/9084.png  \n",
            "   creating: data/train/ID2/1/\n",
            "  inflating: data/train/ID2/1/10069.png  \n",
            "  inflating: data/train/ID2/1/11419.png  \n",
            "  inflating: data/train/ID2/1/16695.png  \n",
            "  inflating: data/train/ID2/1/21392.png  \n",
            "  inflating: data/train/ID2/1/21553.png  \n",
            " extracting: data/train/ID2/1/24689.png  \n",
            " extracting: data/train/ID2/1/27331.png  \n",
            " extracting: data/train/ID2/1/44328.png  \n",
            " extracting: data/train/ID2/1/58690.png  \n",
            " extracting: data/train/ID2/1/6373.png  \n",
            "   creating: data/train/ID2/2/\n",
            " extracting: data/train/ID2/2/17449.png  \n",
            " extracting: data/train/ID2/2/19192.png  \n",
            " extracting: data/train/ID2/2/2252.png  \n",
            " extracting: data/train/ID2/2/26469.png  \n",
            "  inflating: data/train/ID2/2/27958.png  \n",
            " extracting: data/train/ID2/2/2980.png  \n",
            " extracting: data/train/ID2/2/31493.png  \n",
            " extracting: data/train/ID2/2/3713.png  \n",
            " extracting: data/train/ID2/2/39098.png  \n",
            " extracting: data/train/ID2/2/48935.png  \n",
            "   creating: data/train/ID2/3/\n",
            "   creating: data/train/ID2/4/\n",
            " extracting: data/train/ID2/4/10671.png  \n",
            " extracting: data/train/ID2/4/12956.png  \n",
            " extracting: data/train/ID2/4/1843.png  \n",
            " extracting: data/train/ID2/4/19854.png  \n",
            "  inflating: data/train/ID2/4/24317.png  \n",
            " extracting: data/train/ID2/4/31404.png  \n",
            " extracting: data/train/ID2/4/33286.png  \n",
            " extracting: data/train/ID2/4/34949.png  \n",
            "  inflating: data/train/ID2/4/39792.png  \n",
            " extracting: data/train/ID2/4/54246.png  \n",
            "   creating: data/train/ID2/5/\n",
            " extracting: data/train/ID2/5/15013.png  \n",
            "  inflating: data/train/ID2/5/16575.png  \n",
            " extracting: data/train/ID2/5/16833.png  \n",
            " extracting: data/train/ID2/5/19084.png  \n",
            " extracting: data/train/ID2/5/23939.png  \n",
            " extracting: data/train/ID2/5/3267.png  \n",
            " extracting: data/train/ID2/5/33045.png  \n",
            " extracting: data/train/ID2/5/35868.png  \n",
            " extracting: data/train/ID2/5/5178.png  \n",
            " extracting: data/train/ID2/5/54751.png  \n",
            "   creating: data/train/ID2/6/\n",
            " extracting: data/train/ID2/6/14769.png  \n",
            " extracting: data/train/ID2/6/26766.png  \n",
            " extracting: data/train/ID2/6/28681.png  \n",
            " extracting: data/train/ID2/6/31229.png  \n",
            " extracting: data/train/ID2/6/45139.png  \n",
            " extracting: data/train/ID2/6/51805.png  \n",
            " extracting: data/train/ID2/6/54206.png  \n",
            " extracting: data/train/ID2/6/5767.png  \n",
            " extracting: data/train/ID2/6/59812.png  \n",
            " extracting: data/train/ID2/6/763.png  \n",
            "   creating: data/train/ID2/7/\n",
            " extracting: data/train/ID2/7/11787.png  \n",
            " extracting: data/train/ID2/7/14440.png  \n",
            " extracting: data/train/ID2/7/16849.png  \n",
            " extracting: data/train/ID2/7/39878.png  \n",
            " extracting: data/train/ID2/7/49739.png  \n",
            "  inflating: data/train/ID2/7/52580.png  \n",
            " extracting: data/train/ID2/7/53511.png  \n",
            " extracting: data/train/ID2/7/54386.png  \n",
            " extracting: data/train/ID2/7/57673.png  \n",
            " extracting: data/train/ID2/7/7060.png  \n",
            "   creating: data/train/ID2/8/\n",
            " extracting: data/train/ID2/8/11312.png  \n",
            " extracting: data/train/ID2/8/21689.png  \n",
            " extracting: data/train/ID2/8/33669.png  \n",
            " extracting: data/train/ID2/8/37006.png  \n",
            " extracting: data/train/ID2/8/39648.png  \n",
            " extracting: data/train/ID2/8/48163.png  \n",
            " extracting: data/train/ID2/8/49752.png  \n",
            " extracting: data/train/ID2/8/54353.png  \n",
            " extracting: data/train/ID2/8/8723.png  \n",
            " extracting: data/train/ID2/8/9695.png  \n",
            "   creating: data/train/ID2/9/\n",
            " extracting: data/train/ID2/9/23749.png  \n",
            " extracting: data/train/ID2/9/24157.png  \n",
            " extracting: data/train/ID2/9/26847.png  \n",
            " extracting: data/train/ID2/9/31219.png  \n",
            " extracting: data/train/ID2/9/3529.png  \n",
            " extracting: data/train/ID2/9/3650.png  \n",
            " extracting: data/train/ID2/9/45776.png  \n",
            " extracting: data/train/ID2/9/48385.png  \n",
            " extracting: data/train/ID2/9/55494.png  \n",
            " extracting: data/train/ID2/9/56409.png  \n",
            "   creating: data/train/ID3/\n",
            "   creating: data/train/ID3/0/\n",
            " extracting: data/train/ID3/0/12800.png  \n",
            " extracting: data/train/ID3/0/20206.png  \n",
            " extracting: data/train/ID3/0/21243.png  \n",
            " extracting: data/train/ID3/0/2840.png  \n",
            " extracting: data/train/ID3/0/31678.png  \n",
            " extracting: data/train/ID3/0/34353.png  \n",
            " extracting: data/train/ID3/0/40822.png  \n",
            " extracting: data/train/ID3/0/50128.png  \n",
            "  inflating: data/train/ID3/0/51465.png  \n",
            " extracting: data/train/ID3/0/59144.png  \n",
            "   creating: data/train/ID3/1/\n",
            " extracting: data/train/ID3/1/10469.png  \n",
            " extracting: data/train/ID3/1/11680.png  \n",
            "  inflating: data/train/ID3/1/15153.png  \n",
            "  inflating: data/train/ID3/1/26803.png  \n",
            " extracting: data/train/ID3/1/27620.png  \n",
            " extracting: data/train/ID3/1/30750.png  \n",
            "  inflating: data/train/ID3/1/34427.png  \n",
            " extracting: data/train/ID3/1/39019.png  \n",
            " extracting: data/train/ID3/1/52370.png  \n",
            "  inflating: data/train/ID3/1/9481.png  \n",
            "   creating: data/train/ID3/2/\n",
            "  inflating: data/train/ID3/2/1125.png  \n",
            " extracting: data/train/ID3/2/15108.png  \n",
            " extracting: data/train/ID3/2/29655.png  \n",
            " extracting: data/train/ID3/2/36716.png  \n",
            " extracting: data/train/ID3/2/39511.png  \n",
            " extracting: data/train/ID3/2/40852.png  \n",
            " extracting: data/train/ID3/2/52371.png  \n",
            " extracting: data/train/ID3/2/52756.png  \n",
            " extracting: data/train/ID3/2/54433.png  \n",
            " extracting: data/train/ID3/2/59476.png  \n",
            "   creating: data/train/ID3/3/\n",
            " extracting: data/train/ID3/3/1433.png  \n",
            " extracting: data/train/ID3/3/17156.png  \n",
            " extracting: data/train/ID3/3/20081.png  \n",
            " extracting: data/train/ID3/3/26695.png  \n",
            " extracting: data/train/ID3/3/33897.png  \n",
            "  inflating: data/train/ID3/3/37066.png  \n",
            " extracting: data/train/ID3/3/37637.png  \n",
            " extracting: data/train/ID3/3/56261.png  \n",
            " extracting: data/train/ID3/3/7583.png  \n",
            "  inflating: data/train/ID3/3/7591.png  \n",
            "   creating: data/train/ID3/4/\n",
            " extracting: data/train/ID3/4/12078.png  \n",
            " extracting: data/train/ID3/4/22243.png  \n",
            " extracting: data/train/ID3/4/22772.png  \n",
            "  inflating: data/train/ID3/4/29136.png  \n",
            " extracting: data/train/ID3/4/29575.png  \n",
            "  inflating: data/train/ID3/4/36641.png  \n",
            " extracting: data/train/ID3/4/45319.png  \n",
            " extracting: data/train/ID3/4/46921.png  \n",
            " extracting: data/train/ID3/4/49275.png  \n",
            "  inflating: data/train/ID3/4/7735.png  \n",
            "   creating: data/train/ID3/5/\n",
            "   creating: data/train/ID3/6/\n",
            " extracting: data/train/ID3/6/17981.png  \n",
            " extracting: data/train/ID3/6/23285.png  \n",
            " extracting: data/train/ID3/6/42526.png  \n",
            " extracting: data/train/ID3/6/44910.png  \n",
            " extracting: data/train/ID3/6/47220.png  \n",
            " extracting: data/train/ID3/6/4923.png  \n",
            " extracting: data/train/ID3/6/50161.png  \n",
            " extracting: data/train/ID3/6/59792.png  \n",
            " extracting: data/train/ID3/6/7009.png  \n",
            " extracting: data/train/ID3/6/817.png  \n",
            "   creating: data/train/ID3/7/\n",
            "  inflating: data/train/ID3/7/14077.png  \n",
            " extracting: data/train/ID3/7/1587.png  \n",
            " extracting: data/train/ID3/7/17799.png  \n",
            " extracting: data/train/ID3/7/33051.png  \n",
            " extracting: data/train/ID3/7/39255.png  \n",
            " extracting: data/train/ID3/7/40379.png  \n",
            " extracting: data/train/ID3/7/46154.png  \n",
            " extracting: data/train/ID3/7/55519.png  \n",
            " extracting: data/train/ID3/7/9225.png  \n",
            " extracting: data/train/ID3/7/9989.png  \n",
            "   creating: data/train/ID3/8/\n",
            " extracting: data/train/ID3/8/14792.png  \n",
            " extracting: data/train/ID3/8/19001.png  \n",
            " extracting: data/train/ID3/8/24647.png  \n",
            " extracting: data/train/ID3/8/30546.png  \n",
            " extracting: data/train/ID3/8/30782.png  \n",
            " extracting: data/train/ID3/8/36817.png  \n",
            " extracting: data/train/ID3/8/44222.png  \n",
            " extracting: data/train/ID3/8/51088.png  \n",
            " extracting: data/train/ID3/8/53421.png  \n",
            " extracting: data/train/ID3/8/59098.png  \n",
            "   creating: data/train/ID3/9/\n",
            " extracting: data/train/ID3/9/10795.png  \n",
            "  inflating: data/train/ID3/9/16847.png  \n",
            " extracting: data/train/ID3/9/21572.png  \n",
            " extracting: data/train/ID3/9/27345.png  \n",
            " extracting: data/train/ID3/9/35388.png  \n",
            " extracting: data/train/ID3/9/38494.png  \n",
            " extracting: data/train/ID3/9/4631.png  \n",
            " extracting: data/train/ID3/9/6669.png  \n",
            " extracting: data/train/ID3/9/7512.png  \n",
            " extracting: data/train/ID3/9/9950.png  \n",
            "   creating: data/train/ID4/\n",
            "   creating: data/train/ID4/0/\n",
            " extracting: data/train/ID4/0/11218.png  \n",
            " extracting: data/train/ID4/0/15383.png  \n",
            " extracting: data/train/ID4/0/19858.png  \n",
            " extracting: data/train/ID4/0/26340.png  \n",
            " extracting: data/train/ID4/0/36214.png  \n",
            " extracting: data/train/ID4/0/41558.png  \n",
            " extracting: data/train/ID4/0/47740.png  \n",
            " extracting: data/train/ID4/0/50750.png  \n",
            " extracting: data/train/ID4/0/56675.png  \n",
            " extracting: data/train/ID4/0/59198.png  \n",
            "   creating: data/train/ID4/1/\n",
            " extracting: data/train/ID4/1/24184.png  \n",
            " extracting: data/train/ID4/1/26582.png  \n",
            " extracting: data/train/ID4/1/30431.png  \n",
            " extracting: data/train/ID4/1/34763.png  \n",
            " extracting: data/train/ID4/1/37434.png  \n",
            "  inflating: data/train/ID4/1/37850.png  \n",
            " extracting: data/train/ID4/1/41817.png  \n",
            " extracting: data/train/ID4/1/46108.png  \n",
            "  inflating: data/train/ID4/1/47735.png  \n",
            " extracting: data/train/ID4/1/48664.png  \n",
            "   creating: data/train/ID4/2/\n",
            " extracting: data/train/ID4/2/21136.png  \n",
            "  inflating: data/train/ID4/2/28959.png  \n",
            " extracting: data/train/ID4/2/30753.png  \n",
            " extracting: data/train/ID4/2/38382.png  \n",
            " extracting: data/train/ID4/2/39007.png  \n",
            " extracting: data/train/ID4/2/49499.png  \n",
            " extracting: data/train/ID4/2/52657.png  \n",
            " extracting: data/train/ID4/2/55988.png  \n",
            "  inflating: data/train/ID4/2/57443.png  \n",
            " extracting: data/train/ID4/2/6191.png  \n",
            "   creating: data/train/ID4/3/\n",
            " extracting: data/train/ID4/3/11202.png  \n",
            " extracting: data/train/ID4/3/1676.png  \n",
            " extracting: data/train/ID4/3/21999.png  \n",
            " extracting: data/train/ID4/3/36263.png  \n",
            " extracting: data/train/ID4/3/48726.png  \n",
            "  inflating: data/train/ID4/3/51246.png  \n",
            " extracting: data/train/ID4/3/52228.png  \n",
            " extracting: data/train/ID4/3/52546.png  \n",
            " extracting: data/train/ID4/3/53756.png  \n",
            "  inflating: data/train/ID4/3/59981.png  \n",
            "   creating: data/train/ID4/4/\n",
            " extracting: data/train/ID4/4/23993.png  \n",
            " extracting: data/train/ID4/4/25972.png  \n",
            "  inflating: data/train/ID4/4/30299.png  \n",
            " extracting: data/train/ID4/4/32179.png  \n",
            " extracting: data/train/ID4/4/35760.png  \n",
            " extracting: data/train/ID4/4/48195.png  \n",
            " extracting: data/train/ID4/4/50183.png  \n",
            " extracting: data/train/ID4/4/51078.png  \n",
            " extracting: data/train/ID4/4/52025.png  \n",
            " extracting: data/train/ID4/4/8581.png  \n",
            "   creating: data/train/ID4/5/\n",
            " extracting: data/train/ID4/5/19853.png  \n",
            " extracting: data/train/ID4/5/20141.png  \n",
            " extracting: data/train/ID4/5/21866.png  \n",
            " extracting: data/train/ID4/5/23837.png  \n",
            " extracting: data/train/ID4/5/39723.png  \n",
            "  inflating: data/train/ID4/5/45433.png  \n",
            " extracting: data/train/ID4/5/50297.png  \n",
            "  inflating: data/train/ID4/5/51252.png  \n",
            " extracting: data/train/ID4/5/6170.png  \n",
            " extracting: data/train/ID4/5/6408.png  \n",
            "   creating: data/train/ID4/6/\n",
            "   creating: data/train/ID4/7/\n",
            " extracting: data/train/ID4/7/10719.png  \n",
            " extracting: data/train/ID4/7/10990.png  \n",
            " extracting: data/train/ID4/7/16402.png  \n",
            " extracting: data/train/ID4/7/18258.png  \n",
            " extracting: data/train/ID4/7/25427.png  \n",
            " extracting: data/train/ID4/7/27633.png  \n",
            " extracting: data/train/ID4/7/30505.png  \n",
            " extracting: data/train/ID4/7/42825.png  \n",
            " extracting: data/train/ID4/7/48222.png  \n",
            "  inflating: data/train/ID4/7/48520.png  \n",
            "   creating: data/train/ID4/8/\n",
            " extracting: data/train/ID4/8/31309.png  \n",
            " extracting: data/train/ID4/8/35429.png  \n",
            " extracting: data/train/ID4/8/36914.png  \n",
            " extracting: data/train/ID4/8/37629.png  \n",
            " extracting: data/train/ID4/8/37769.png  \n",
            " extracting: data/train/ID4/8/38086.png  \n",
            " extracting: data/train/ID4/8/38864.png  \n",
            " extracting: data/train/ID4/8/41305.png  \n",
            "  inflating: data/train/ID4/8/48705.png  \n",
            " extracting: data/train/ID4/8/6621.png  \n",
            "   creating: data/train/ID4/9/\n",
            " extracting: data/train/ID4/9/17182.png  \n",
            " extracting: data/train/ID4/9/1906.png  \n",
            " extracting: data/train/ID4/9/19587.png  \n",
            " extracting: data/train/ID4/9/2723.png  \n",
            " extracting: data/train/ID4/9/28985.png  \n",
            " extracting: data/train/ID4/9/3415.png  \n",
            " extracting: data/train/ID4/9/35192.png  \n",
            " extracting: data/train/ID4/9/36008.png  \n",
            " extracting: data/train/ID4/9/45233.png  \n",
            " extracting: data/train/ID4/9/59567.png  \n",
            "   creating: data/train/ID5/\n",
            "   creating: data/train/ID5/0/\n",
            " extracting: data/train/ID5/0/1503.png  \n",
            " extracting: data/train/ID5/0/18030.png  \n",
            " extracting: data/train/ID5/0/27736.png  \n",
            " extracting: data/train/ID5/0/33238.png  \n",
            " extracting: data/train/ID5/0/33391.png  \n",
            " extracting: data/train/ID5/0/3462.png  \n",
            " extracting: data/train/ID5/0/34868.png  \n",
            " extracting: data/train/ID5/0/40238.png  \n",
            " extracting: data/train/ID5/0/43206.png  \n",
            " extracting: data/train/ID5/0/44697.png  \n",
            "   creating: data/train/ID5/1/\n",
            "  inflating: data/train/ID5/1/12972.png  \n",
            " extracting: data/train/ID5/1/14057.png  \n",
            "  inflating: data/train/ID5/1/1473.png  \n",
            " extracting: data/train/ID5/1/18408.png  \n",
            " extracting: data/train/ID5/1/19672.png  \n",
            " extracting: data/train/ID5/1/37379.png  \n",
            " extracting: data/train/ID5/1/41104.png  \n",
            " extracting: data/train/ID5/1/42562.png  \n",
            " extracting: data/train/ID5/1/45963.png  \n",
            "  inflating: data/train/ID5/1/9619.png  \n",
            "   creating: data/train/ID5/2/\n",
            " extracting: data/train/ID5/2/11272.png  \n",
            "  inflating: data/train/ID5/2/19464.png  \n",
            " extracting: data/train/ID5/2/22779.png  \n",
            " extracting: data/train/ID5/2/30832.png  \n",
            " extracting: data/train/ID5/2/40367.png  \n",
            " extracting: data/train/ID5/2/40719.png  \n",
            " extracting: data/train/ID5/2/41205.png  \n",
            " extracting: data/train/ID5/2/42444.png  \n",
            " extracting: data/train/ID5/2/56786.png  \n",
            " extracting: data/train/ID5/2/59975.png  \n",
            "   creating: data/train/ID5/3/\n",
            "   creating: data/train/ID5/4/\n",
            " extracting: data/train/ID5/4/10372.png  \n",
            " extracting: data/train/ID5/4/13072.png  \n",
            " extracting: data/train/ID5/4/17493.png  \n",
            "  inflating: data/train/ID5/4/1923.png  \n",
            "  inflating: data/train/ID5/4/19984.png  \n",
            " extracting: data/train/ID5/4/25163.png  \n",
            "  inflating: data/train/ID5/4/35850.png  \n",
            " extracting: data/train/ID5/4/43813.png  \n",
            " extracting: data/train/ID5/4/4758.png  \n",
            " extracting: data/train/ID5/4/59535.png  \n",
            "   creating: data/train/ID5/5/\n",
            " extracting: data/train/ID5/5/17100.png  \n",
            " extracting: data/train/ID5/5/23641.png  \n",
            " extracting: data/train/ID5/5/23747.png  \n",
            " extracting: data/train/ID5/5/27779.png  \n",
            " extracting: data/train/ID5/5/32687.png  \n",
            " extracting: data/train/ID5/5/4105.png  \n",
            " extracting: data/train/ID5/5/43839.png  \n",
            " extracting: data/train/ID5/5/49784.png  \n",
            " extracting: data/train/ID5/5/6087.png  \n",
            " extracting: data/train/ID5/5/9647.png  \n",
            "   creating: data/train/ID5/6/\n",
            " extracting: data/train/ID5/6/11986.png  \n",
            " extracting: data/train/ID5/6/23020.png  \n",
            " extracting: data/train/ID5/6/28468.png  \n",
            " extracting: data/train/ID5/6/39252.png  \n",
            " extracting: data/train/ID5/6/42396.png  \n",
            " extracting: data/train/ID5/6/45550.png  \n",
            " extracting: data/train/ID5/6/50824.png  \n",
            " extracting: data/train/ID5/6/50844.png  \n",
            " extracting: data/train/ID5/6/57778.png  \n",
            " extracting: data/train/ID5/6/7582.png  \n",
            "   creating: data/train/ID5/7/\n",
            "  inflating: data/train/ID5/7/14077.png  \n",
            " extracting: data/train/ID5/7/16195.png  \n",
            " extracting: data/train/ID5/7/18143.png  \n",
            "  inflating: data/train/ID5/7/23150.png  \n",
            "  inflating: data/train/ID5/7/2660.png  \n",
            " extracting: data/train/ID5/7/27668.png  \n",
            " extracting: data/train/ID5/7/31883.png  \n",
            " extracting: data/train/ID5/7/34862.png  \n",
            "  inflating: data/train/ID5/7/44006.png  \n",
            "  inflating: data/train/ID5/7/4726.png  \n",
            "   creating: data/train/ID5/8/\n",
            " extracting: data/train/ID5/8/10890.png  \n",
            " extracting: data/train/ID5/8/12887.png  \n",
            " extracting: data/train/ID5/8/13190.png  \n",
            " extracting: data/train/ID5/8/19666.png  \n",
            " extracting: data/train/ID5/8/23988.png  \n",
            " extracting: data/train/ID5/8/29432.png  \n",
            " extracting: data/train/ID5/8/34987.png  \n",
            " extracting: data/train/ID5/8/38743.png  \n",
            " extracting: data/train/ID5/8/39480.png  \n",
            " extracting: data/train/ID5/8/8137.png  \n",
            "   creating: data/train/ID5/9/\n",
            " extracting: data/train/ID5/9/15554.png  \n",
            " extracting: data/train/ID5/9/16471.png  \n",
            " extracting: data/train/ID5/9/25638.png  \n",
            " extracting: data/train/ID5/9/27227.png  \n",
            " extracting: data/train/ID5/9/30309.png  \n",
            "  inflating: data/train/ID5/9/33423.png  \n",
            " extracting: data/train/ID5/9/35764.png  \n",
            "  inflating: data/train/ID5/9/37061.png  \n",
            " extracting: data/train/ID5/9/40532.png  \n",
            " extracting: data/train/ID5/9/41590.png  \n",
            "   creating: data/train/ID6/\n",
            "   creating: data/train/ID6/0/\n",
            " extracting: data/train/ID6/0/1269.png  \n",
            " extracting: data/train/ID6/0/13372.png  \n",
            " extracting: data/train/ID6/0/21359.png  \n",
            " extracting: data/train/ID6/0/46055.png  \n",
            " extracting: data/train/ID6/0/55049.png  \n",
            " extracting: data/train/ID6/0/57065.png  \n",
            " extracting: data/train/ID6/0/57327.png  \n",
            " extracting: data/train/ID6/0/57687.png  \n",
            " extracting: data/train/ID6/0/6445.png  \n",
            " extracting: data/train/ID6/0/7272.png  \n",
            "   creating: data/train/ID6/1/\n",
            "  inflating: data/train/ID6/1/10099.png  \n",
            " extracting: data/train/ID6/1/12072.png  \n",
            "  inflating: data/train/ID6/1/1666.png  \n",
            " extracting: data/train/ID6/1/23281.png  \n",
            " extracting: data/train/ID6/1/29616.png  \n",
            " extracting: data/train/ID6/1/33280.png  \n",
            " extracting: data/train/ID6/1/39556.png  \n",
            " extracting: data/train/ID6/1/43390.png  \n",
            " extracting: data/train/ID6/1/7717.png  \n",
            " extracting: data/train/ID6/1/8088.png  \n",
            "   creating: data/train/ID6/2/\n",
            " extracting: data/train/ID6/2/25736.png  \n",
            " extracting: data/train/ID6/2/29785.png  \n",
            " extracting: data/train/ID6/2/38222.png  \n",
            " extracting: data/train/ID6/2/41541.png  \n",
            " extracting: data/train/ID6/2/41917.png  \n",
            " extracting: data/train/ID6/2/44299.png  \n",
            " extracting: data/train/ID6/2/46664.png  \n",
            " extracting: data/train/ID6/2/49268.png  \n",
            " extracting: data/train/ID6/2/53133.png  \n",
            " extracting: data/train/ID6/2/55382.png  \n",
            "   creating: data/train/ID6/3/\n",
            " extracting: data/train/ID6/3/13525.png  \n",
            " extracting: data/train/ID6/3/30516.png  \n",
            " extracting: data/train/ID6/3/42399.png  \n",
            " extracting: data/train/ID6/3/43760.png  \n",
            " extracting: data/train/ID6/3/43869.png  \n",
            " extracting: data/train/ID6/3/51999.png  \n",
            " extracting: data/train/ID6/3/52905.png  \n",
            " extracting: data/train/ID6/3/55051.png  \n",
            " extracting: data/train/ID6/3/56198.png  \n",
            "  inflating: data/train/ID6/3/6907.png  \n",
            "   creating: data/train/ID6/4/\n",
            " extracting: data/train/ID6/4/18055.png  \n",
            " extracting: data/train/ID6/4/20908.png  \n",
            " extracting: data/train/ID6/4/25094.png  \n",
            " extracting: data/train/ID6/4/28656.png  \n",
            " extracting: data/train/ID6/4/35636.png  \n",
            " extracting: data/train/ID6/4/36201.png  \n",
            " extracting: data/train/ID6/4/48985.png  \n",
            " extracting: data/train/ID6/4/52280.png  \n",
            "  inflating: data/train/ID6/4/52349.png  \n",
            " extracting: data/train/ID6/4/58464.png  \n",
            "   creating: data/train/ID6/5/\n",
            " extracting: data/train/ID6/5/15612.png  \n",
            " extracting: data/train/ID6/5/2714.png  \n",
            " extracting: data/train/ID6/5/2756.png  \n",
            " extracting: data/train/ID6/5/32672.png  \n",
            " extracting: data/train/ID6/5/35030.png  \n",
            " extracting: data/train/ID6/5/36154.png  \n",
            " extracting: data/train/ID6/5/39231.png  \n",
            " extracting: data/train/ID6/5/43128.png  \n",
            " extracting: data/train/ID6/5/46942.png  \n",
            " extracting: data/train/ID6/5/58376.png  \n",
            "   creating: data/train/ID6/6/\n",
            " extracting: data/train/ID6/6/10726.png  \n",
            " extracting: data/train/ID6/6/11685.png  \n",
            " extracting: data/train/ID6/6/20958.png  \n",
            " extracting: data/train/ID6/6/37991.png  \n",
            " extracting: data/train/ID6/6/39165.png  \n",
            " extracting: data/train/ID6/6/45756.png  \n",
            "  inflating: data/train/ID6/6/49085.png  \n",
            " extracting: data/train/ID6/6/50267.png  \n",
            " extracting: data/train/ID6/6/59471.png  \n",
            " extracting: data/train/ID6/6/7969.png  \n",
            "   creating: data/train/ID6/7/\n",
            "   creating: data/train/ID6/8/\n",
            " extracting: data/train/ID6/8/15051.png  \n",
            " extracting: data/train/ID6/8/16023.png  \n",
            " extracting: data/train/ID6/8/24890.png  \n",
            " extracting: data/train/ID6/8/34820.png  \n",
            " extracting: data/train/ID6/8/35265.png  \n",
            " extracting: data/train/ID6/8/38864.png  \n",
            " extracting: data/train/ID6/8/46873.png  \n",
            " extracting: data/train/ID6/8/47310.png  \n",
            " extracting: data/train/ID6/8/51144.png  \n",
            " extracting: data/train/ID6/8/52466.png  \n",
            "   creating: data/train/ID6/9/\n",
            " extracting: data/train/ID6/9/11993.png  \n",
            "  inflating: data/train/ID6/9/20827.png  \n",
            " extracting: data/train/ID6/9/23503.png  \n",
            " extracting: data/train/ID6/9/29592.png  \n",
            " extracting: data/train/ID6/9/3360.png  \n",
            " extracting: data/train/ID6/9/34666.png  \n",
            " extracting: data/train/ID6/9/38457.png  \n",
            " extracting: data/train/ID6/9/43826.png  \n",
            "  inflating: data/train/ID6/9/50529.png  \n",
            " extracting: data/train/ID6/9/57130.png  \n",
            "   creating: data/train/ID7/\n",
            "   creating: data/train/ID7/0/\n",
            " extracting: data/train/ID7/0/11847.png  \n",
            " extracting: data/train/ID7/0/15629.png  \n",
            " extracting: data/train/ID7/0/25513.png  \n",
            " extracting: data/train/ID7/0/3744.png  \n",
            " extracting: data/train/ID7/0/38356.png  \n",
            " extracting: data/train/ID7/0/42363.png  \n",
            " extracting: data/train/ID7/0/45113.png  \n",
            " extracting: data/train/ID7/0/8528.png  \n",
            " extracting: data/train/ID7/0/9553.png  \n",
            " extracting: data/train/ID7/0/9923.png  \n",
            "   creating: data/train/ID7/1/\n",
            " extracting: data/train/ID7/1/17926.png  \n",
            " extracting: data/train/ID7/1/18410.png  \n",
            " extracting: data/train/ID7/1/22476.png  \n",
            " extracting: data/train/ID7/1/23890.png  \n",
            " extracting: data/train/ID7/1/30090.png  \n",
            "  inflating: data/train/ID7/1/30149.png  \n",
            " extracting: data/train/ID7/1/37615.png  \n",
            " extracting: data/train/ID7/1/50208.png  \n",
            " extracting: data/train/ID7/1/51950.png  \n",
            " extracting: data/train/ID7/1/58308.png  \n",
            "   creating: data/train/ID7/2/\n",
            "   creating: data/train/ID7/3/\n",
            " extracting: data/train/ID7/3/12896.png  \n",
            " extracting: data/train/ID7/3/21150.png  \n",
            " extracting: data/train/ID7/3/24736.png  \n",
            " extracting: data/train/ID7/3/26831.png  \n",
            " extracting: data/train/ID7/3/36296.png  \n",
            " extracting: data/train/ID7/3/4401.png  \n",
            " extracting: data/train/ID7/3/45879.png  \n",
            " extracting: data/train/ID7/3/57489.png  \n",
            " extracting: data/train/ID7/3/57834.png  \n",
            " extracting: data/train/ID7/3/8499.png  \n",
            "   creating: data/train/ID7/4/\n",
            " extracting: data/train/ID7/4/10968.png  \n",
            " extracting: data/train/ID7/4/11693.png  \n",
            " extracting: data/train/ID7/4/14135.png  \n",
            "  inflating: data/train/ID7/4/17391.png  \n",
            " extracting: data/train/ID7/4/19402.png  \n",
            " extracting: data/train/ID7/4/25671.png  \n",
            " extracting: data/train/ID7/4/30780.png  \n",
            "  inflating: data/train/ID7/4/41441.png  \n",
            "  inflating: data/train/ID7/4/43399.png  \n",
            " extracting: data/train/ID7/4/57748.png  \n",
            "   creating: data/train/ID7/5/\n",
            "  inflating: data/train/ID7/5/19220.png  \n",
            " extracting: data/train/ID7/5/22157.png  \n",
            "  inflating: data/train/ID7/5/23192.png  \n",
            " extracting: data/train/ID7/5/37050.png  \n",
            " extracting: data/train/ID7/5/43504.png  \n",
            " extracting: data/train/ID7/5/43588.png  \n",
            "  inflating: data/train/ID7/5/43734.png  \n",
            " extracting: data/train/ID7/5/47846.png  \n",
            " extracting: data/train/ID7/5/53051.png  \n",
            "  inflating: data/train/ID7/5/54132.png  \n",
            "   creating: data/train/ID7/6/\n",
            " extracting: data/train/ID7/6/1267.png  \n",
            " extracting: data/train/ID7/6/1394.png  \n",
            " extracting: data/train/ID7/6/23126.png  \n",
            " extracting: data/train/ID7/6/32842.png  \n",
            " extracting: data/train/ID7/6/35628.png  \n",
            " extracting: data/train/ID7/6/42848.png  \n",
            " extracting: data/train/ID7/6/43438.png  \n",
            " extracting: data/train/ID7/6/43683.png  \n",
            " extracting: data/train/ID7/6/45751.png  \n",
            " extracting: data/train/ID7/6/56864.png  \n",
            "   creating: data/train/ID7/7/\n",
            " extracting: data/train/ID7/7/19345.png  \n",
            " extracting: data/train/ID7/7/22531.png  \n",
            " extracting: data/train/ID7/7/26055.png  \n",
            "  inflating: data/train/ID7/7/36053.png  \n",
            " extracting: data/train/ID7/7/46082.png  \n",
            " extracting: data/train/ID7/7/50795.png  \n",
            "  inflating: data/train/ID7/7/51651.png  \n",
            " extracting: data/train/ID7/7/5368.png  \n",
            " extracting: data/train/ID7/7/5627.png  \n",
            " extracting: data/train/ID7/7/57926.png  \n",
            "   creating: data/train/ID7/8/\n",
            " extracting: data/train/ID7/8/12624.png  \n",
            " extracting: data/train/ID7/8/14453.png  \n",
            " extracting: data/train/ID7/8/2483.png  \n",
            " extracting: data/train/ID7/8/33387.png  \n",
            " extracting: data/train/ID7/8/38310.png  \n",
            " extracting: data/train/ID7/8/39916.png  \n",
            " extracting: data/train/ID7/8/44790.png  \n",
            " extracting: data/train/ID7/8/50525.png  \n",
            " extracting: data/train/ID7/8/55853.png  \n",
            " extracting: data/train/ID7/8/6729.png  \n",
            "   creating: data/train/ID7/9/\n",
            " extracting: data/train/ID7/9/1353.png  \n",
            " extracting: data/train/ID7/9/17922.png  \n",
            " extracting: data/train/ID7/9/2007.png  \n",
            " extracting: data/train/ID7/9/30104.png  \n",
            " extracting: data/train/ID7/9/32418.png  \n",
            " extracting: data/train/ID7/9/33644.png  \n",
            " extracting: data/train/ID7/9/48567.png  \n",
            " extracting: data/train/ID7/9/51241.png  \n",
            "  inflating: data/train/ID7/9/54579.png  \n",
            " extracting: data/train/ID7/9/58128.png  \n",
            "   creating: data/train/ID8/\n",
            "   creating: data/train/ID8/0/\n",
            " extracting: data/train/ID8/0/15972.png  \n",
            " extracting: data/train/ID8/0/21050.png  \n",
            " extracting: data/train/ID8/0/24859.png  \n",
            " extracting: data/train/ID8/0/28889.png  \n",
            " extracting: data/train/ID8/0/33258.png  \n",
            " extracting: data/train/ID8/0/33678.png  \n",
            " extracting: data/train/ID8/0/41042.png  \n",
            " extracting: data/train/ID8/0/54273.png  \n",
            " extracting: data/train/ID8/0/5712.png  \n",
            " extracting: data/train/ID8/0/57812.png  \n",
            "   creating: data/train/ID8/1/\n",
            " extracting: data/train/ID8/1/1201.png  \n",
            " extracting: data/train/ID8/1/12167.png  \n",
            "  inflating: data/train/ID8/1/2363.png  \n",
            " extracting: data/train/ID8/1/29144.png  \n",
            " extracting: data/train/ID8/1/2933.png  \n",
            " extracting: data/train/ID8/1/33417.png  \n",
            " extracting: data/train/ID8/1/46934.png  \n",
            " extracting: data/train/ID8/1/55084.png  \n",
            " extracting: data/train/ID8/1/57006.png  \n",
            " extracting: data/train/ID8/1/6710.png  \n",
            "   creating: data/train/ID8/2/\n",
            " extracting: data/train/ID8/2/19398.png  \n",
            " extracting: data/train/ID8/2/24090.png  \n",
            " extracting: data/train/ID8/2/3660.png  \n",
            " extracting: data/train/ID8/2/42742.png  \n",
            " extracting: data/train/ID8/2/42959.png  \n",
            "  inflating: data/train/ID8/2/45853.png  \n",
            "  inflating: data/train/ID8/2/46552.png  \n",
            " extracting: data/train/ID8/2/48641.png  \n",
            " extracting: data/train/ID8/2/55382.png  \n",
            " extracting: data/train/ID8/2/55920.png  \n",
            "   creating: data/train/ID8/3/\n",
            " extracting: data/train/ID8/3/11959.png  \n",
            " extracting: data/train/ID8/3/12968.png  \n",
            " extracting: data/train/ID8/3/13907.png  \n",
            " extracting: data/train/ID8/3/16020.png  \n",
            " extracting: data/train/ID8/3/27032.png  \n",
            " extracting: data/train/ID8/3/37941.png  \n",
            " extracting: data/train/ID8/3/43903.png  \n",
            " extracting: data/train/ID8/3/45656.png  \n",
            " extracting: data/train/ID8/3/46131.png  \n",
            " extracting: data/train/ID8/3/59047.png  \n",
            "   creating: data/train/ID8/4/\n",
            "  inflating: data/train/ID8/4/1634.png  \n",
            " extracting: data/train/ID8/4/17897.png  \n",
            "  inflating: data/train/ID8/4/23275.png  \n",
            "  inflating: data/train/ID8/4/3281.png  \n",
            " extracting: data/train/ID8/4/37959.png  \n",
            " extracting: data/train/ID8/4/43027.png  \n",
            " extracting: data/train/ID8/4/43689.png  \n",
            " extracting: data/train/ID8/4/47501.png  \n",
            " extracting: data/train/ID8/4/53348.png  \n",
            " extracting: data/train/ID8/4/57418.png  \n",
            "   creating: data/train/ID8/5/\n",
            " extracting: data/train/ID8/5/10537.png  \n",
            " extracting: data/train/ID8/5/10702.png  \n",
            " extracting: data/train/ID8/5/35275.png  \n",
            " extracting: data/train/ID8/5/37181.png  \n",
            " extracting: data/train/ID8/5/39514.png  \n",
            "  inflating: data/train/ID8/5/41322.png  \n",
            " extracting: data/train/ID8/5/52510.png  \n",
            " extracting: data/train/ID8/5/5255.png  \n",
            " extracting: data/train/ID8/5/6982.png  \n",
            " extracting: data/train/ID8/5/7846.png  \n",
            "   creating: data/train/ID8/6/\n",
            "  inflating: data/train/ID8/6/12403.png  \n",
            " extracting: data/train/ID8/6/19452.png  \n",
            " extracting: data/train/ID8/6/29769.png  \n",
            " extracting: data/train/ID8/6/37832.png  \n",
            "  inflating: data/train/ID8/6/40395.png  \n",
            "  inflating: data/train/ID8/6/40758.png  \n",
            " extracting: data/train/ID8/6/40850.png  \n",
            " extracting: data/train/ID8/6/48329.png  \n",
            "  inflating: data/train/ID8/6/50344.png  \n",
            " extracting: data/train/ID8/6/55040.png  \n",
            "   creating: data/train/ID8/7/\n",
            " extracting: data/train/ID8/7/1218.png  \n",
            " extracting: data/train/ID8/7/21463.png  \n",
            " extracting: data/train/ID8/7/22863.png  \n",
            " extracting: data/train/ID8/7/33308.png  \n",
            " extracting: data/train/ID8/7/34752.png  \n",
            "  inflating: data/train/ID8/7/35197.png  \n",
            " extracting: data/train/ID8/7/35447.png  \n",
            " extracting: data/train/ID8/7/37395.png  \n",
            " extracting: data/train/ID8/7/50047.png  \n",
            "  inflating: data/train/ID8/7/57335.png  \n",
            "   creating: data/train/ID8/8/\n",
            "  inflating: data/train/ID8/8/17629.png  \n",
            " extracting: data/train/ID8/8/22809.png  \n",
            " extracting: data/train/ID8/8/24298.png  \n",
            " extracting: data/train/ID8/8/25987.png  \n",
            " extracting: data/train/ID8/8/2811.png  \n",
            " extracting: data/train/ID8/8/3439.png  \n",
            " extracting: data/train/ID8/8/4007.png  \n",
            " extracting: data/train/ID8/8/49702.png  \n",
            " extracting: data/train/ID8/8/59396.png  \n",
            " extracting: data/train/ID8/8/721.png  \n",
            "   creating: data/train/ID8/9/\n",
            "   creating: data/train/ID9/\n",
            "   creating: data/train/ID9/0/\n",
            " extracting: data/train/ID9/0/17188.png  \n",
            " extracting: data/train/ID9/0/2587.png  \n",
            " extracting: data/train/ID9/0/28467.png  \n",
            " extracting: data/train/ID9/0/29395.png  \n",
            " extracting: data/train/ID9/0/34052.png  \n",
            " extracting: data/train/ID9/0/36332.png  \n",
            " extracting: data/train/ID9/0/40442.png  \n",
            " extracting: data/train/ID9/0/42447.png  \n",
            " extracting: data/train/ID9/0/51599.png  \n",
            " extracting: data/train/ID9/0/7272.png  \n",
            "   creating: data/train/ID9/1/\n",
            " extracting: data/train/ID9/1/16715.png  \n",
            "  inflating: data/train/ID9/1/17908.png  \n",
            "  inflating: data/train/ID9/1/27346.png  \n",
            "  inflating: data/train/ID9/1/36118.png  \n",
            "  inflating: data/train/ID9/1/45007.png  \n",
            "  inflating: data/train/ID9/1/4904.png  \n",
            " extracting: data/train/ID9/1/5175.png  \n",
            " extracting: data/train/ID9/1/6464.png  \n",
            " extracting: data/train/ID9/1/7403.png  \n",
            " extracting: data/train/ID9/1/8352.png  \n",
            "   creating: data/train/ID9/2/\n",
            " extracting: data/train/ID9/2/1144.png  \n",
            " extracting: data/train/ID9/2/23562.png  \n",
            " extracting: data/train/ID9/2/27163.png  \n",
            " extracting: data/train/ID9/2/30832.png  \n",
            " extracting: data/train/ID9/2/3201.png  \n",
            " extracting: data/train/ID9/2/39562.png  \n",
            " extracting: data/train/ID9/2/47259.png  \n",
            " extracting: data/train/ID9/2/50954.png  \n",
            " extracting: data/train/ID9/2/55986.png  \n",
            " extracting: data/train/ID9/2/6479.png  \n",
            "   creating: data/train/ID9/3/\n",
            "   creating: data/train/ID9/4/\n",
            " extracting: data/train/ID9/4/11284.png  \n",
            " extracting: data/train/ID9/4/12949.png  \n",
            " extracting: data/train/ID9/4/14958.png  \n",
            "  inflating: data/train/ID9/4/21917.png  \n",
            " extracting: data/train/ID9/4/23993.png  \n",
            " extracting: data/train/ID9/4/28502.png  \n",
            " extracting: data/train/ID9/4/31115.png  \n",
            " extracting: data/train/ID9/4/46504.png  \n",
            " extracting: data/train/ID9/4/52159.png  \n",
            " extracting: data/train/ID9/4/55038.png  \n",
            "   creating: data/train/ID9/5/\n",
            "  inflating: data/train/ID9/5/12918.png  \n",
            " extracting: data/train/ID9/5/16981.png  \n",
            " extracting: data/train/ID9/5/21125.png  \n",
            " extracting: data/train/ID9/5/27001.png  \n",
            " extracting: data/train/ID9/5/2913.png  \n",
            " extracting: data/train/ID9/5/39973.png  \n",
            " extracting: data/train/ID9/5/40994.png  \n",
            "  inflating: data/train/ID9/5/45592.png  \n",
            " extracting: data/train/ID9/5/54812.png  \n",
            " extracting: data/train/ID9/5/56033.png  \n",
            "   creating: data/train/ID9/6/\n",
            " extracting: data/train/ID9/6/11254.png  \n",
            " extracting: data/train/ID9/6/16119.png  \n",
            " extracting: data/train/ID9/6/40597.png  \n",
            " extracting: data/train/ID9/6/51379.png  \n",
            " extracting: data/train/ID9/6/5375.png  \n",
            " extracting: data/train/ID9/6/55383.png  \n",
            " extracting: data/train/ID9/6/57558.png  \n",
            " extracting: data/train/ID9/6/57803.png  \n",
            " extracting: data/train/ID9/6/58167.png  \n",
            " extracting: data/train/ID9/6/7196.png  \n",
            "   creating: data/train/ID9/7/\n",
            " extracting: data/train/ID9/7/14506.png  \n",
            " extracting: data/train/ID9/7/14834.png  \n",
            " extracting: data/train/ID9/7/17008.png  \n",
            " extracting: data/train/ID9/7/19227.png  \n",
            " extracting: data/train/ID9/7/26273.png  \n",
            " extracting: data/train/ID9/7/3864.png  \n",
            " extracting: data/train/ID9/7/42379.png  \n",
            " extracting: data/train/ID9/7/45646.png  \n",
            "  inflating: data/train/ID9/7/5913.png  \n",
            " extracting: data/train/ID9/7/7421.png  \n",
            "   creating: data/train/ID9/8/\n",
            " extracting: data/train/ID9/8/1389.png  \n",
            " extracting: data/train/ID9/8/23769.png  \n",
            " extracting: data/train/ID9/8/2385.png  \n",
            " extracting: data/train/ID9/8/25430.png  \n",
            " extracting: data/train/ID9/8/35243.png  \n",
            " extracting: data/train/ID9/8/42300.png  \n",
            " extracting: data/train/ID9/8/46739.png  \n",
            " extracting: data/train/ID9/8/48724.png  \n",
            " extracting: data/train/ID9/8/54296.png  \n",
            " extracting: data/train/ID9/8/54742.png  \n",
            "   creating: data/train/ID9/9/\n",
            " extracting: data/train/ID9/9/12739.png  \n",
            " extracting: data/train/ID9/9/28320.png  \n",
            "  inflating: data/train/ID9/9/29092.png  \n",
            " extracting: data/train/ID9/9/30506.png  \n",
            " extracting: data/train/ID9/9/34340.png  \n",
            " extracting: data/train/ID9/9/36457.png  \n",
            " extracting: data/train/ID9/9/43739.png  \n",
            " extracting: data/train/ID9/9/44045.png  \n",
            " extracting: data/train/ID9/9/52844.png  \n",
            " extracting: data/train/ID9/9/7812.png  \n"
          ]
        }
      ],
      "source": [
        "!unzip /content/drive/MyDrive/minist_overfit.zip"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "vDgmKvcwFoSZ"
      },
      "source": [
        "Importing libraries"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 20,
      "metadata": {
        "id": "AbR3SNeoFdnV"
      },
      "outputs": [],
      "source": [
        "import numpy as np\n",
        "import pandas as pd\n",
        "import matplotlib.pyplot as plt\n",
        "import matplotlib.image as img\n",
        "import torch\n",
        "import torch.nn as nn # All neural network modules, nn.Linear, nn.Conv2d, BatchNorm, Loss functions\n",
        "import torchvision.datasets as datasets # Has standard datasets we can import in a nice way\n",
        "import torchvision.transforms as transforms # Transformations we can perform on our dataset\n",
        "import torch.nn.functional as F # All functions that don't have any parameters\n",
        "from torch.utils.data import DataLoader, Dataset # Gives easier dataset managment and creates mini batches\n",
        "import torch.optim as optim # For all Optimization algorithms, SGD, Adam, etc.\n",
        "#from torchvision.transforms import ToTensor, Resize, Normalize, Compose, InterpolationMode\n",
        "import math\n",
        "\n",
        "%matplotlib inline\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "tjJlHx-eHL8M"
      },
      "source": [
        "dataset class"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 83,
      "metadata": {
        "id": "uSDkiXhJBa5n"
      },
      "outputs": [],
      "source": [
        "import os\n",
        "from PIL import Image\n",
        "import glob\n",
        "\n",
        "class dataset(Dataset):\n",
        "    \n",
        "    def __init__(self, data_path, transform=None):\n",
        "        self.data_path = data_path\n",
        "        self.transform = transform\n",
        "\n",
        "        self.images = sorted(glob.glob(os.path.join(data_path, '*','*', '*.png')))\n",
        "        \n",
        "        # self.data = []\n",
        "        # self.targets = []\n",
        "        \n",
        "        # # Load data and targets\n",
        "        # for target_class in os.listdir(data_path):\n",
        "        #     target_path = os.path.join(data_path, target_class)\n",
        "        #     target_id = int(target_class)\n",
        "            \n",
        "        #     for image_file in os.listdir(target_path):\n",
        "        #         image_path = os.path.join(target_path, image_file)\n",
        "        #         image = Image.open(image_path)\n",
        "                \n",
        "        #         self.data.append(image)\n",
        "        #         self.targets.append(target_id)\n",
        "        \n",
        "        # self.data = np.array(self.data)\n",
        "        # self.targets = np.array(self.targets)\n",
        "        \n",
        "    def __len__(self):\n",
        "        return len(self.images)\n",
        "    \n",
        "    def __getitem__(self, index):\n",
        "        image_path = self.images[index]\n",
        "        image = Image.open(image_path)\n",
        "        \n",
        "        if self.transform is not None:\n",
        "            image = self.transform(image)\n",
        "        \n",
        "        target = int(os.path.basename(os.path.dirname(image_path)))\n",
        "        \n",
        "        return image, target\n"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "a = glob.glob(os.path.join('/content/data/train', '*','*', '*.png'))\n",
        "a[0:10]"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "NZ7pUY4WXXcT",
        "outputId": "938873cb-8875-4fa7-c3f8-36b765aef185"
      },
      "execution_count": 79,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "['/content/data/train/ID8/1/2933.png',\n",
              " '/content/data/train/ID8/1/29144.png',\n",
              " '/content/data/train/ID8/1/33417.png',\n",
              " '/content/data/train/ID8/1/46934.png',\n",
              " '/content/data/train/ID8/1/12167.png',\n",
              " '/content/data/train/ID8/1/2363.png',\n",
              " '/content/data/train/ID8/1/1201.png',\n",
              " '/content/data/train/ID8/1/6710.png',\n",
              " '/content/data/train/ID8/1/55084.png',\n",
              " '/content/data/train/ID8/1/57006.png']"
            ]
          },
          "metadata": {},
          "execution_count": 79
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "wRvOmlqSIz9-"
      },
      "source": [
        "1. Data loading and normalization"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 53,
      "metadata": {
        "id": "m7xq72E_I8sF"
      },
      "outputs": [],
      "source": [
        "def load_dataset(train_data_path: str, train_size, test_data_path: str,  batch_Size: int, Shuffle: bool):\n",
        "  # Defining transforms for the training data and testing data\n",
        "\n",
        "  train_transform = transforms.Compose([transforms.Resize((28,28)),\n",
        "                                        transforms.Grayscale(num_output_channels=1),\n",
        "                                        transforms.ToTensor(),\n",
        "                                        transforms.Normalize(mean=[0.0], std=[0.5])])\n",
        "  \n",
        "  test_transform = transforms.Compose([transforms.Resize((28,28)),\n",
        "                                       transforms.Grayscale(num_output_channels=1),\n",
        "                                       transforms.ToTensor()])\n",
        "  \n",
        "  \n",
        "  # Loading dataset using ImageFolder method. Passing transforms in here and using test transform for validation data.\n",
        "\n",
        "  train_data = dataset(train_data_path, transform=train_transform)\n",
        "  test_data = dataset(test_data_path, transform=test_transform)\n",
        "\n",
        "  size = len(train_data)\n",
        "  val_size = 1 - train_size\n",
        "\n",
        "  train_data, val_data = torch.utils.data.random_split(train_data, [math.ceil(size * train_size),\n",
        "                                                                    math.ceil(size * val_size)])\n",
        "\n",
        "  \n",
        "  # Optional case if validation data is given\n",
        "  validation_transform = transforms.Compose([transforms.Resize((28,28)),\n",
        "                                             transforms.Grayscale(num_output_channels=1),\n",
        "                                             transforms.ToTensor()])\n",
        "  \n",
        "  #validation_data = datasets.ImageFolder(val_data, transform=validation_transform)\n",
        "  validation_loader = DataLoader(val_data, batch_size=batch_Size)\n",
        "\n",
        "  train_loader = DataLoader(train_data, batch_size=batch_Size)\n",
        "  test_loader = DataLoader(test_data)\n",
        "\n",
        "  print(\"Training Images : \",len(train_data), \"Validation Images\",len(val_data), \"Testing Images\",len(test_data))\n",
        "  # Return \n",
        "  return train_loader, test_loader, validation_loader\n"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "trans = transforms.Compose([transforms.ToTensor()])\n",
        "train = dataset('/content/data/train', transform=trans)\n",
        "train_loader = DataLoader(train,batch_size=10, shuffle=True)\n",
        "train_ = iter(train_loader)"
      ],
      "metadata": {
        "id": "OibejpXkZZkX"
      },
      "execution_count": 98,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "next(train_)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "H7F1PYMibbHp",
        "outputId": "ca7a2f6a-ecdf-41a9-c1f0-00cb14a327c0"
      },
      "execution_count": 99,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[tensor([[[[0., 0., 0.,  ..., 0., 0., 0.],\n",
              "           [0., 0., 0.,  ..., 0., 0., 0.],\n",
              "           [0., 0., 0.,  ..., 0., 0., 0.],\n",
              "           ...,\n",
              "           [0., 0., 0.,  ..., 0., 0., 0.],\n",
              "           [0., 0., 0.,  ..., 0., 0., 0.],\n",
              "           [0., 0., 0.,  ..., 0., 0., 0.]]],\n",
              " \n",
              " \n",
              "         [[[0., 0., 0.,  ..., 0., 0., 0.],\n",
              "           [0., 0., 0.,  ..., 0., 0., 0.],\n",
              "           [0., 0., 0.,  ..., 0., 0., 0.],\n",
              "           ...,\n",
              "           [0., 0., 0.,  ..., 0., 0., 0.],\n",
              "           [0., 0., 0.,  ..., 0., 0., 0.],\n",
              "           [0., 0., 0.,  ..., 0., 0., 0.]]],\n",
              " \n",
              " \n",
              "         [[[0., 0., 0.,  ..., 0., 0., 0.],\n",
              "           [0., 0., 0.,  ..., 0., 0., 0.],\n",
              "           [0., 0., 0.,  ..., 0., 0., 0.],\n",
              "           ...,\n",
              "           [0., 0., 0.,  ..., 0., 0., 0.],\n",
              "           [0., 0., 0.,  ..., 0., 0., 0.],\n",
              "           [0., 0., 0.,  ..., 0., 0., 0.]]],\n",
              " \n",
              " \n",
              "         ...,\n",
              " \n",
              " \n",
              "         [[[0., 0., 0.,  ..., 0., 0., 0.],\n",
              "           [0., 0., 0.,  ..., 0., 0., 0.],\n",
              "           [0., 0., 0.,  ..., 0., 0., 0.],\n",
              "           ...,\n",
              "           [0., 0., 0.,  ..., 0., 0., 0.],\n",
              "           [0., 0., 0.,  ..., 0., 0., 0.],\n",
              "           [0., 0., 0.,  ..., 0., 0., 0.]]],\n",
              " \n",
              " \n",
              "         [[[0., 0., 0.,  ..., 0., 0., 0.],\n",
              "           [0., 0., 0.,  ..., 0., 0., 0.],\n",
              "           [0., 0., 0.,  ..., 0., 0., 0.],\n",
              "           ...,\n",
              "           [0., 0., 0.,  ..., 0., 0., 0.],\n",
              "           [0., 0., 0.,  ..., 0., 0., 0.],\n",
              "           [0., 0., 0.,  ..., 0., 0., 0.]]],\n",
              " \n",
              " \n",
              "         [[[0., 0., 0.,  ..., 0., 0., 0.],\n",
              "           [0., 0., 0.,  ..., 0., 0., 0.],\n",
              "           [0., 0., 0.,  ..., 0., 0., 0.],\n",
              "           ...,\n",
              "           [0., 0., 0.,  ..., 0., 0., 0.],\n",
              "           [0., 0., 0.,  ..., 0., 0., 0.],\n",
              "           [0., 0., 0.,  ..., 0., 0., 0.]]]]),\n",
              " tensor([9, 8, 9, 5, 7, 8, 4, 6, 1, 2])]"
            ]
          },
          "metadata": {},
          "execution_count": 99
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "qon1teDuLAxc"
      },
      "source": [
        "2. Declaring Neural Network"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 6,
      "metadata": {
        "id": "uc3UQsf8PIeB"
      },
      "outputs": [],
      "source": [
        "class Model(nn.Module):\n",
        "    def __init__(self, no_of_layers, input_dim, neurons_per_layer, dropout):\n",
        "        super().__init__()\n",
        "\n",
        "        self.fc_in = nn.Linear(input_dim, neurons_per_layer[0])\n",
        "        self.layers = nn.ModuleList()\n",
        "        \n",
        "        for i in range(no_of_layers - 1):\n",
        "            self.layers.append(nn.Linear(neurons_per_layer[i], neurons_per_layer[i+1]))\n",
        "\n",
        "        self.fc_out = nn.Linear(neurons_per_layer[-1], 10)\n",
        "\n",
        "        self.dropout = nn.Dropout(p=dropout)\n",
        "\n",
        "    def forward(self, x):\n",
        "        x = x.view(x.size(0), -1)\n",
        "        x = self.dropout(F.relu(self.fc_in(x)))\n",
        "        for j in self.layers:\n",
        "            x = self.dropout(F.relu(j(x)))\n",
        "        x = self.fc_out(x)\n",
        "\n",
        "        return x\n",
        "\n",
        "  "
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "f311AtNcToB-"
      },
      "source": [
        "3. Training"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 59,
      "metadata": {
        "id": "ql4c4sYgTmMB"
      },
      "outputs": [],
      "source": [
        "def train(net, train_loader, valid_loader, epochs, loss_func, optimizer, device, visualize):\n",
        "    train_loss_array = []\n",
        "    train_acc_array = []\n",
        "    validation_acc_array = []\n",
        "    validation_loss_array = []\n",
        "    prev_valid_acc = 0.0\n",
        "    \n",
        "    for epoch in range(epochs):\n",
        "        net.train()\n",
        "        train_loss = 0.0\n",
        "        train_correct = 0\n",
        "        train_total = 0\n",
        "        valid_correct = 0\n",
        "        valid_total = 0\n",
        "\n",
        "        for batch_idx, (data, target) in enumerate(train_loader):\n",
        "            data, target = data.to(device), target.to(device)\n",
        "            #print(data[0].shape)\n",
        "            optimizer.zero_grad()\n",
        "            output = net(data)\n",
        "            loss = loss_func(output, target)\n",
        "            loss.backward()\n",
        "            optimizer.step()\n",
        "\n",
        "            # calculate training accuracy\n",
        "            predicted = torch.argmax(output, dim=1)\n",
        "            train_total += target.size(0)\n",
        "            train_correct += (predicted == target).sum().item()\n",
        "\n",
        "            train_loss += loss.item()\n",
        "\n",
        "        train_accuracy = train_correct / train_total\n",
        "        train_acc_array.append(train_accuracy)\n",
        "\n",
        "        # calculate validation accuracy\n",
        "        net.eval()\n",
        "        with torch.no_grad():\n",
        "            valid_loss = 0.0\n",
        "            for data, target in valid_loader:\n",
        "                data, target = data.to(device), target.to(device)\n",
        "                output = net(data)\n",
        "                predicted = torch.argmax(output, dim=1)\n",
        "                valid_total += target.size(0)\n",
        "                valid_correct += (predicted == target).sum().item()\n",
        "                loss = loss_func(output, target)\n",
        "                valid_loss += loss.item()\n",
        "\n",
        "            valid_accuracy = valid_correct / valid_total\n",
        "            validation_acc_array.append(valid_accuracy)\n",
        "\n",
        "            valid_los = valid_loss / len(valid_loader)\n",
        "            validation_loss_array.append(valid_los)\n",
        "\n",
        "            # save the model with current epoch number as the model name\n",
        "            if valid_accuracy > prev_valid_acc:\n",
        "                save_model(net, f'model_{epoch+1}')\n",
        "                prev_valid_acc = valid_accuracy\n",
        "\n",
        "\n",
        "\n",
        "        epoch_loss = train_loss / len(train_loader)\n",
        "        train_loss_array.append(epoch_loss)\n",
        "\n",
        "        print('Epoch {}/{}  Train Loss: {:.6f}  Train Acc: {:.6f}  Valid Loss: {:.6f}  Valid Acc: {:.6f}'.format(\n",
        "            epoch+1, epochs, epoch_loss, train_accuracy, valid_loss/len(valid_loader), valid_accuracy))\n",
        "    if visualize:\n",
        "        train_visualize(train_loss_array=train_loss_array, train_accuracy_array=train_acc_array,\n",
        "                        val_loss_array=validation_loss_array,\n",
        "                        val_accuracy_array=validation_acc_array)\n",
        "\n",
        "        \n",
        "    return net, train_loss_array, train_acc_array, validation_loss_array, validation_acc_array\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "u43vjP1MI9bT"
      },
      "source": [
        "4. Save Network"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 8,
      "metadata": {
        "id": "fxjSgF-e50qv"
      },
      "outputs": [],
      "source": [
        "def save_model(model, model_name: str):\n",
        "    torch.save(model, '{}.pt'.format(model_name))\n",
        "    "
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "inh90WBMJmTC"
      },
      "source": [
        "5. Load Network"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 9,
      "metadata": {
        "id": "eiNSwYmbJefp"
      },
      "outputs": [],
      "source": [
        "def load_model(model_file_path: str):\n",
        "    return torch.load(model_file_path)\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "APOQsxkmKLkz"
      },
      "source": [
        "6. Testing Step\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 10,
      "metadata": {
        "id": "-Ny7l2qLJ-NV"
      },
      "outputs": [],
      "source": [
        "def test(model, test_loader,loss_func, visualize):\n",
        "    with torch.no_grad():\n",
        "        pred = []\n",
        "        output = []\n",
        "        data_ = test_loader.dataset\n",
        "        loss_array = []\n",
        "        for index, (data, batch_output) in enumerate(test_loader):\n",
        "            out = model(data)\n",
        "            pred_ = torch.argmax(out, dim=1)\n",
        "            pred.append(pred_)\n",
        "            output.append(batch_output)\n",
        "            loss = loss_func(out, batch_output)\n",
        "            loss_array.append(loss.item())\n",
        "\n",
        "        pred = torch.cat(pred, dim=0)\n",
        "        output = torch.cat(output, dim=0)\n",
        "\n",
        "        plt.plot(range(len(loss_array)), loss_array)\n",
        "        plt.xlabel('no of samples')\n",
        "        plt.ylabel('loss')\n",
        "        plt.title('Loss curve for test losses')\n",
        "        plt.show()\n",
        "        \n",
        "        \n",
        "        correct = (pred == output).sum().item()\n",
        "        total = output.size(0)\n",
        "        accuracy = correct/total\n",
        "\n",
        "    if visualize:\n",
        "        test_visualize(pred_output=pred, true_output=output, data=data_, accuracy=accuracy)\n",
        "    return pred\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "OeGbA9ieZ-1z"
      },
      "source": [
        "7. Visualize Results\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 72,
      "metadata": {
        "id": "6LlZUY-9NESq"
      },
      "outputs": [],
      "source": [
        "from sklearn.metrics import f1_score, confusion_matrix, classification_report\n",
        "import seaborn as sns\n",
        "def train_visualize(train_loss_array, train_accuracy_array, val_loss_array, val_accuracy_array):\n",
        "    fig, ax = plt.subplots(1, 2, figsize=(20, 5))\n",
        "    ax[0].plot(train_loss_array, color='royalblue', label='Train Loss')\n",
        "    ax[0].plot(val_loss_array, color='darkred', label='Validation Loss')\n",
        "    ax[0].set_xlabel('NO. of epochs')\n",
        "    ax[0].set_ylabel('Loss')\n",
        "    ax[0].set_title('Training and Validation Losses')\n",
        "    ax[0].legend()\n",
        "\n",
        "    ax[1].plot(train_accuracy_array, color='royalblue', label='Train Accuracy')\n",
        "    ax[1].plot(val_accuracy_array, color='darkred', label='Validation Accuracy')\n",
        "    ax[1].set_xlabel('NO. of epochs')\n",
        "    ax[1].set_ylabel('Accuracy')\n",
        "    ax[1].set_title('Training and Validation Accuracy')\n",
        "    ax[1].legend()\n",
        "    plt.show()\n",
        "\n",
        "def test_visualize(pred_output, true_output, data, accuracy):\n",
        "    n = 0\n",
        "    for i in range(len(true_output)):\n",
        "        if pred_output[i] != true_output[i] and n < 5:\n",
        "            img = data[i][0].cpu().numpy().squeeze()  # convert to NumPy array and remove channel dimension\n",
        "            plt.imshow(img, cmap='gray')\n",
        "            plt.xlabel('Predicted output is: {}'.format(pred_output[i]))\n",
        "            plt.show()\n",
        "            n += 1\n",
        "\n",
        "        if i%500 == 0:\n",
        "            img = data[i][0].cpu().numpy().squeeze()  # convert to NumPy array and remove channel dimension\n",
        "            plt.imshow(img, cmap='gray')\n",
        "            plt.xlabel('Predicted output is: {}'.format(pred_output[i]))\n",
        "            plt.show()\n",
        "\n",
        "    Confusion_matrix = confusion_matrix(true_output, pred_output)\n",
        "    classification = classification_report(true_output, pred_output)\n",
        "\n",
        "    fig = plt.figure(figsize=(15,5))\n",
        "    sns.heatmap(Confusion_matrix, annot=True, cmap='Blues')\n",
        "    plt.show()\n",
        "\n",
        "    \n",
        "    print('Classification_report is: \\n {}\\n Accuracy is: {}'.format(classification, accuracy))\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "iROfRcpp5NkX"
      },
      "source": [
        "8. Main Function"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 12,
      "metadata": {
        "id": "k2A7M9K9nHW0"
      },
      "outputs": [],
      "source": [
        "def main_function(training_path, testing_path, train_size, input_dim, no_of_hidden_layers,\n",
        "                  list_of_neurons, optimizer, learning_rate, no_of_epchos, loss_function,\n",
        "                  batch_size, dropout, Gpu=False, is_model=False, Model_path=None,\n",
        "                  visualize_results=False, save_Model=False, model_name=None):\n",
        "    if Gpu and torch.cuda.is_available():\n",
        "        device = torch.device('cuda')\n",
        "    else:\n",
        "        device = torch.device('cpu')\n",
        "\n",
        "    # Loading data sets\n",
        "    train_loader, test_loader, validation_loader = load_dataset(training_path, train_size,\n",
        "                                                                testing_path,\n",
        "                                                                batch_size,\n",
        "                                                                Shuffle=True)\n",
        "\n",
        "    # training\n",
        "    if is_model:\n",
        "        model = load_model(Model_path)\n",
        "    else:\n",
        "        model = Model(no_of_hidden_layers, input_dim=input_dim, neurons_per_layer=list_of_neurons, dropout=dropout)\n",
        "        \n",
        "    \n",
        "    optimizer1 = optimizer(model.parameters(), lr=learning_rate)\n",
        "    model, train_loss_array, train_accuracy_array,validation_loss_array, validation_acc_array = train(net= model,\n",
        "                                                                                            train_loader=train_loader,\n",
        "                                                                                          valid_loader=validation_loader,\n",
        "                                                                                           epochs=no_of_epchos,\n",
        "                                                                                            loss_func=loss_function,\n",
        "                                                                                             optimizer=optimizer1,\n",
        "                                                                                              device=device,\n",
        "                                                                                            visualize=visualize_results)\n",
        "    # Testing\n",
        "    prediction = test(model=model, test_loader=test_loader,loss_func=loss_function, visualize=visualize_results)\n",
        "\n",
        "    # Saving model \n",
        "    if save_Model:\n",
        "        save_model(model, model_name=model_name)\n",
        "\n",
        "    \n",
        "    "
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ummwCeZlrAYH"
      },
      "source": [
        "Calling main function"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 71,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "id": "r_v5KBK8rDXV",
        "outputId": "bb67156f-cb32-457c-91ff-30f607c2fc14"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Training Images :  720 Validation Images 180 Testing Images 240\n",
            "Epoch 1/100  Train Loss: 17.705887  Train Acc: 0.118056  Valid Loss: 5.884571  Valid Acc: 0.083333\n",
            "Epoch 2/100  Train Loss: 4.176616  Train Acc: 0.111111  Valid Loss: 2.866570  Valid Acc: 0.088889\n",
            "Epoch 3/100  Train Loss: 2.843892  Train Acc: 0.108333  Valid Loss: 2.547842  Valid Acc: 0.088889\n",
            "Epoch 4/100  Train Loss: 2.552403  Train Acc: 0.112500  Valid Loss: 2.441880  Valid Acc: 0.083333\n",
            "Epoch 5/100  Train Loss: 2.486089  Train Acc: 0.104167  Valid Loss: 2.401871  Valid Acc: 0.100000\n",
            "Epoch 6/100  Train Loss: 2.389960  Train Acc: 0.115278  Valid Loss: 2.374525  Valid Acc: 0.100000\n",
            "Epoch 7/100  Train Loss: 2.352316  Train Acc: 0.109722  Valid Loss: 2.355426  Valid Acc: 0.094444\n",
            "Epoch 8/100  Train Loss: 2.352462  Train Acc: 0.116667  Valid Loss: 2.339657  Valid Acc: 0.094444\n",
            "Epoch 9/100  Train Loss: 2.314076  Train Acc: 0.126389  Valid Loss: 2.336711  Valid Acc: 0.094444\n",
            "Epoch 10/100  Train Loss: 2.307711  Train Acc: 0.115278  Valid Loss: 2.339644  Valid Acc: 0.094444\n",
            "Epoch 11/100  Train Loss: 2.295593  Train Acc: 0.111111  Valid Loss: 2.336071  Valid Acc: 0.094444\n",
            "Epoch 12/100  Train Loss: 2.284082  Train Acc: 0.133333  Valid Loss: 2.332960  Valid Acc: 0.094444\n",
            "Epoch 13/100  Train Loss: 2.262497  Train Acc: 0.131944  Valid Loss: 2.331394  Valid Acc: 0.094444\n",
            "Epoch 14/100  Train Loss: 2.257695  Train Acc: 0.152778  Valid Loss: 2.333959  Valid Acc: 0.094444\n",
            "Epoch 15/100  Train Loss: 2.268860  Train Acc: 0.137500  Valid Loss: 2.337672  Valid Acc: 0.072222\n",
            "Epoch 16/100  Train Loss: 2.255485  Train Acc: 0.144444  Valid Loss: 2.341842  Valid Acc: 0.072222\n",
            "Epoch 17/100  Train Loss: 2.245647  Train Acc: 0.141667  Valid Loss: 2.338590  Valid Acc: 0.072222\n",
            "Epoch 18/100  Train Loss: 2.252229  Train Acc: 0.148611  Valid Loss: 2.346287  Valid Acc: 0.072222\n",
            "Epoch 19/100  Train Loss: 2.237720  Train Acc: 0.140278  Valid Loss: 2.342973  Valid Acc: 0.077778\n",
            "Epoch 20/100  Train Loss: 2.231773  Train Acc: 0.138889  Valid Loss: 2.351131  Valid Acc: 0.083333\n",
            "Epoch 21/100  Train Loss: 2.225263  Train Acc: 0.152778  Valid Loss: 2.350941  Valid Acc: 0.083333\n",
            "Epoch 22/100  Train Loss: 2.209933  Train Acc: 0.168056  Valid Loss: 2.356340  Valid Acc: 0.072222\n",
            "Epoch 23/100  Train Loss: 2.191930  Train Acc: 0.177778  Valid Loss: 2.364795  Valid Acc: 0.061111\n",
            "Epoch 24/100  Train Loss: 2.195690  Train Acc: 0.150000  Valid Loss: 2.382730  Valid Acc: 0.066667\n",
            "Epoch 25/100  Train Loss: 2.166773  Train Acc: 0.176389  Valid Loss: 2.388247  Valid Acc: 0.050000\n",
            "Epoch 26/100  Train Loss: 2.174779  Train Acc: 0.175000  Valid Loss: 2.385028  Valid Acc: 0.066667\n",
            "Epoch 27/100  Train Loss: 2.150070  Train Acc: 0.176389  Valid Loss: 2.392787  Valid Acc: 0.050000\n",
            "Epoch 28/100  Train Loss: 2.122782  Train Acc: 0.191667  Valid Loss: 2.393370  Valid Acc: 0.055556\n",
            "Epoch 29/100  Train Loss: 2.151880  Train Acc: 0.180556  Valid Loss: 2.395719  Valid Acc: 0.055556\n",
            "Epoch 30/100  Train Loss: 2.149699  Train Acc: 0.163889  Valid Loss: 2.393621  Valid Acc: 0.038889\n",
            "Epoch 31/100  Train Loss: 2.089886  Train Acc: 0.209722  Valid Loss: 2.409260  Valid Acc: 0.044444\n",
            "Epoch 32/100  Train Loss: 2.091310  Train Acc: 0.198611  Valid Loss: 2.426364  Valid Acc: 0.050000\n",
            "Epoch 33/100  Train Loss: 2.084526  Train Acc: 0.209722  Valid Loss: 2.437277  Valid Acc: 0.055556\n",
            "Epoch 34/100  Train Loss: 2.074389  Train Acc: 0.238889  Valid Loss: 2.449130  Valid Acc: 0.055556\n",
            "Epoch 35/100  Train Loss: 2.030599  Train Acc: 0.236111  Valid Loss: 2.466383  Valid Acc: 0.055556\n",
            "Epoch 36/100  Train Loss: 2.003562  Train Acc: 0.227778  Valid Loss: 2.479606  Valid Acc: 0.061111\n",
            "Epoch 37/100  Train Loss: 2.032927  Train Acc: 0.234722  Valid Loss: 2.488743  Valid Acc: 0.050000\n",
            "Epoch 38/100  Train Loss: 1.996209  Train Acc: 0.248611  Valid Loss: 2.509716  Valid Acc: 0.055556\n",
            "Epoch 39/100  Train Loss: 1.964358  Train Acc: 0.270833  Valid Loss: 2.530956  Valid Acc: 0.072222\n",
            "Epoch 40/100  Train Loss: 1.977348  Train Acc: 0.247222  Valid Loss: 2.574207  Valid Acc: 0.055556\n",
            "Epoch 41/100  Train Loss: 1.955963  Train Acc: 0.273611  Valid Loss: 2.558054  Valid Acc: 0.055556\n",
            "Epoch 42/100  Train Loss: 1.902706  Train Acc: 0.283333  Valid Loss: 2.575818  Valid Acc: 0.050000\n",
            "Epoch 43/100  Train Loss: 1.905656  Train Acc: 0.273611  Valid Loss: 2.658885  Valid Acc: 0.077778\n",
            "Epoch 44/100  Train Loss: 1.869677  Train Acc: 0.291667  Valid Loss: 2.658660  Valid Acc: 0.072222\n",
            "Epoch 45/100  Train Loss: 1.863549  Train Acc: 0.325000  Valid Loss: 2.625896  Valid Acc: 0.072222\n",
            "Epoch 46/100  Train Loss: 1.868270  Train Acc: 0.312500  Valid Loss: 2.706430  Valid Acc: 0.072222\n",
            "Epoch 47/100  Train Loss: 1.824478  Train Acc: 0.315278  Valid Loss: 2.744049  Valid Acc: 0.072222\n",
            "Epoch 48/100  Train Loss: 1.793467  Train Acc: 0.329167  Valid Loss: 2.752945  Valid Acc: 0.077778\n",
            "Epoch 49/100  Train Loss: 1.776465  Train Acc: 0.330556  Valid Loss: 2.758329  Valid Acc: 0.072222\n",
            "Epoch 50/100  Train Loss: 1.750554  Train Acc: 0.340278  Valid Loss: 2.786947  Valid Acc: 0.066667\n",
            "Epoch 51/100  Train Loss: 1.746434  Train Acc: 0.351389  Valid Loss: 2.820005  Valid Acc: 0.066667\n",
            "Epoch 52/100  Train Loss: 1.706865  Train Acc: 0.366667  Valid Loss: 2.873298  Valid Acc: 0.066667\n",
            "Epoch 53/100  Train Loss: 1.684593  Train Acc: 0.373611  Valid Loss: 2.810598  Valid Acc: 0.077778\n",
            "Epoch 54/100  Train Loss: 1.654065  Train Acc: 0.383333  Valid Loss: 2.903600  Valid Acc: 0.072222\n",
            "Epoch 55/100  Train Loss: 1.587290  Train Acc: 0.404167  Valid Loss: 2.919742  Valid Acc: 0.072222\n",
            "Epoch 56/100  Train Loss: 1.578227  Train Acc: 0.433333  Valid Loss: 2.957125  Valid Acc: 0.088889\n",
            "Epoch 57/100  Train Loss: 1.541380  Train Acc: 0.419444  Valid Loss: 3.014066  Valid Acc: 0.072222\n",
            "Epoch 58/100  Train Loss: 1.550393  Train Acc: 0.438889  Valid Loss: 3.099327  Valid Acc: 0.061111\n",
            "Epoch 59/100  Train Loss: 1.539936  Train Acc: 0.426389  Valid Loss: 3.073802  Valid Acc: 0.094444\n",
            "Epoch 60/100  Train Loss: 1.482552  Train Acc: 0.444444  Valid Loss: 3.088128  Valid Acc: 0.077778\n",
            "Epoch 61/100  Train Loss: 1.458060  Train Acc: 0.470833  Valid Loss: 3.080556  Valid Acc: 0.094444\n",
            "Epoch 62/100  Train Loss: 1.445268  Train Acc: 0.455556  Valid Loss: 3.184638  Valid Acc: 0.077778\n",
            "Epoch 63/100  Train Loss: 1.423081  Train Acc: 0.490278  Valid Loss: 3.193612  Valid Acc: 0.100000\n",
            "Epoch 64/100  Train Loss: 1.388741  Train Acc: 0.509722  Valid Loss: 3.169853  Valid Acc: 0.122222\n",
            "Epoch 65/100  Train Loss: 1.346291  Train Acc: 0.488889  Valid Loss: 3.277003  Valid Acc: 0.100000\n",
            "Epoch 66/100  Train Loss: 1.345718  Train Acc: 0.515278  Valid Loss: 3.365962  Valid Acc: 0.100000\n",
            "Epoch 67/100  Train Loss: 1.289831  Train Acc: 0.541667  Valid Loss: 3.458150  Valid Acc: 0.105556\n",
            "Epoch 68/100  Train Loss: 1.297508  Train Acc: 0.538889  Valid Loss: 3.310206  Valid Acc: 0.094444\n",
            "Epoch 69/100  Train Loss: 1.249754  Train Acc: 0.550000  Valid Loss: 3.432408  Valid Acc: 0.094444\n",
            "Epoch 70/100  Train Loss: 1.194966  Train Acc: 0.559722  Valid Loss: 3.586307  Valid Acc: 0.127778\n",
            "Epoch 71/100  Train Loss: 1.197006  Train Acc: 0.577778  Valid Loss: 3.467653  Valid Acc: 0.105556\n",
            "Epoch 72/100  Train Loss: 1.192116  Train Acc: 0.586111  Valid Loss: 3.544897  Valid Acc: 0.105556\n",
            "Epoch 73/100  Train Loss: 1.094605  Train Acc: 0.615278  Valid Loss: 3.730053  Valid Acc: 0.088889\n",
            "Epoch 74/100  Train Loss: 1.121724  Train Acc: 0.598611  Valid Loss: 3.631712  Valid Acc: 0.083333\n",
            "Epoch 75/100  Train Loss: 1.092887  Train Acc: 0.601389  Valid Loss: 3.570283  Valid Acc: 0.077778\n",
            "Epoch 76/100  Train Loss: 1.042772  Train Acc: 0.631944  Valid Loss: 3.745069  Valid Acc: 0.094444\n",
            "Epoch 77/100  Train Loss: 1.060510  Train Acc: 0.625000  Valid Loss: 3.562861  Valid Acc: 0.105556\n",
            "Epoch 78/100  Train Loss: 0.952362  Train Acc: 0.684722  Valid Loss: 3.811268  Valid Acc: 0.088889\n",
            "Epoch 79/100  Train Loss: 0.970628  Train Acc: 0.659722  Valid Loss: 3.787058  Valid Acc: 0.083333\n",
            "Epoch 80/100  Train Loss: 0.907654  Train Acc: 0.677778  Valid Loss: 4.026687  Valid Acc: 0.105556\n",
            "Epoch 81/100  Train Loss: 0.886175  Train Acc: 0.687500  Valid Loss: 4.047087  Valid Acc: 0.088889\n",
            "Epoch 82/100  Train Loss: 0.822692  Train Acc: 0.691667  Valid Loss: 4.149520  Valid Acc: 0.083333\n",
            "Epoch 83/100  Train Loss: 0.861749  Train Acc: 0.698611  Valid Loss: 4.094278  Valid Acc: 0.111111\n",
            "Epoch 84/100  Train Loss: 0.862218  Train Acc: 0.683333  Valid Loss: 4.085864  Valid Acc: 0.100000\n",
            "Epoch 85/100  Train Loss: 0.814428  Train Acc: 0.677778  Valid Loss: 4.013218  Valid Acc: 0.100000\n",
            "Epoch 86/100  Train Loss: 0.816321  Train Acc: 0.726389  Valid Loss: 4.088180  Valid Acc: 0.105556\n",
            "Epoch 87/100  Train Loss: 0.753143  Train Acc: 0.738889  Valid Loss: 4.052704  Valid Acc: 0.111111\n",
            "Epoch 88/100  Train Loss: 0.782744  Train Acc: 0.720833  Valid Loss: 4.116940  Valid Acc: 0.105556\n",
            "Epoch 89/100  Train Loss: 0.763238  Train Acc: 0.741667  Valid Loss: 4.395473  Valid Acc: 0.088889\n",
            "Epoch 90/100  Train Loss: 0.756480  Train Acc: 0.743056  Valid Loss: 4.622451  Valid Acc: 0.094444\n",
            "Epoch 91/100  Train Loss: 0.729779  Train Acc: 0.751389  Valid Loss: 4.565833  Valid Acc: 0.094444\n",
            "Epoch 92/100  Train Loss: 0.722500  Train Acc: 0.743056  Valid Loss: 4.580816  Valid Acc: 0.077778\n",
            "Epoch 93/100  Train Loss: 0.748539  Train Acc: 0.733333  Valid Loss: 4.556219  Valid Acc: 0.088889\n",
            "Epoch 94/100  Train Loss: 0.675121  Train Acc: 0.768056  Valid Loss: 4.577339  Valid Acc: 0.111111\n",
            "Epoch 95/100  Train Loss: 0.634068  Train Acc: 0.772222  Valid Loss: 4.611594  Valid Acc: 0.105556\n",
            "Epoch 96/100  Train Loss: 0.663917  Train Acc: 0.788889  Valid Loss: 4.606387  Valid Acc: 0.088889\n",
            "Epoch 97/100  Train Loss: 0.583819  Train Acc: 0.800000  Valid Loss: 4.780672  Valid Acc: 0.083333\n",
            "Epoch 98/100  Train Loss: 0.574530  Train Acc: 0.805556  Valid Loss: 4.998753  Valid Acc: 0.088889\n",
            "Epoch 99/100  Train Loss: 0.648860  Train Acc: 0.794444  Valid Loss: 4.977625  Valid Acc: 0.083333\n",
            "Epoch 100/100  Train Loss: 0.557990  Train Acc: 0.818056  Valid Loss: 4.972641  Valid Acc: 0.105556\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 1440x360 with 2 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 1080x360 with 2 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Classification_report is: \n",
            "               precision    recall  f1-score   support\n",
            "\n",
            "           0       0.11      0.15      0.12        20\n",
            "           1       0.00      0.00      0.00        20\n",
            "           2       0.29      0.20      0.24        20\n",
            "           3       0.25      0.27      0.26        26\n",
            "           4       0.06      0.04      0.05        23\n",
            "           5       0.12      0.12      0.12        25\n",
            "           6       0.12      0.17      0.14        24\n",
            "           7       0.30      0.22      0.26        27\n",
            "           8       0.24      0.44      0.31        27\n",
            "           9       0.41      0.25      0.31        28\n",
            "\n",
            "    accuracy                           0.20       240\n",
            "   macro avg       0.19      0.19      0.18       240\n",
            "weighted avg       0.20      0.20      0.19       240\n",
            "\n",
            " Accuracy is: 0.19583333333333333\n"
          ]
        }
      ],
      "source": [
        "# #Loss function: Cross-entropy loss\n",
        "# #Optimizer: Adam optimizer\n",
        "# #Learning rate: 0.001-0.01\n",
        "# #Number of epochs: 20-30\n",
        "# #Number of hidden layers: 2-3\n",
        "# #Number of neurons in each hidden layer: 128-256\n",
        "# #batch size: 32-64\n",
        "# #dropout: 0.2-0.5\n",
        "# loss_fun = nn.CrossEntropyLoss()\n",
        "# hiddenlayers = 4\n",
        "# listof_HL = [400,200,100,50]\n",
        "# drop = 0.2\n",
        "# batch = 32\n",
        "# optimizer = torch.optim.Adam\n",
        "# learning = 0.001\n",
        "# epochs = 20\n",
        "\n",
        "# Over Fitting\n",
        "\n",
        "loss_fun = nn.CrossEntropyLoss()\n",
        "hiddenlayers = 1\n",
        "listof_HL = [128]\n",
        "drop = 0.2\n",
        "batch = 32\n",
        "optimizer = torch.optim.Adam\n",
        "learning = 0.001\n",
        "epochs = 100\n",
        "\n",
        "\n",
        "main_function(training_path='/content/data/train',\n",
        "              testing_path='/content/data/test',\n",
        "              train_size=0.8,\n",
        "              input_dim=784,\n",
        "              no_of_hidden_layers=hiddenlayers,\n",
        "              list_of_neurons=listof_HL, optimizer=optimizer,\n",
        "              learning_rate=learning, no_of_epchos=epochs,\n",
        "              loss_function=loss_fun, batch_size=batch,\n",
        "              dropout=drop, is_model=True,\n",
        "              Model_path='/content/drive/MyDrive/model_17.pt',\n",
        "              visualize_results=True)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 54,
      "metadata": {
        "id": "PGjGhS4WQGbV",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "6e036b7e-9697-4ee0-8a1d-ee0a62f2aa05"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Training Images :  720 Validation Images 180 Testing Images 240\n"
          ]
        }
      ],
      "source": [
        "train_loader, test_loader, validation_loader = load_dataset(train_data_path='/content/data/train', train_size=0.8,\n",
        "                                                                test_data_path='/content/data/test',\n",
        "                                                                batch_Size=32,\n",
        "                                                                Shuffle=True)"
      ]
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "-tUM4Y9jvXFd"
      },
      "execution_count": null,
      "outputs": []
    }
  ],
  "metadata": {
    "colab": {
      "provenance": []
    },
    "kernelspec": {
      "display_name": "Python 3",
      "name": "python3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "nbformat": 4,
  "nbformat_minor": 0
}