geam_classification.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import os\n",
    "import torch\n",
    "import torchvision.transforms as transforms\n",
    "from torchvision.datasets import ImageFolder\n",
    "from torch.utils.data import DataLoader\n",
    "import torch.nn as nn\n",
    "import torch.optim as optim\n",
    "from torchvision import models\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 104,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Keeping E:/datasets/datasets_gemstone/test/Alexandrite\\alexandrite_18.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Alexandrite\\alexandrite_28.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Alexandrite\\alexandrite_3.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Alexandrite\\alexandrite_6.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Almandine\\almandine_18.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Almandine\\almandine_28.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Almandine\\almandine_3.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Almandine\\almandine_9.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Amazonite\\amazonite_18.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Amazonite\\amazonite_28.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Amazonite\\amazonite_3.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Amazonite\\amazonite_8.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Amber\\amber_18.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Amber\\amber_28.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Amber\\amber_3.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Amber\\amber_9.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Amethyst\\amethyst_18.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Amethyst\\amethyst_3.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Amethyst\\amethyst_5.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Amethyst\\amethyst_7.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Ametrine\\ametrine_18.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Ametrine\\ametrine_3.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Ametrine\\ametrine_5.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Ametrine\\ametrine_7.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Andalusite\\andalusite_18.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Andalusite\\andalusite_28.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Andalusite\\andalusite_3.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Andalusite\\andalusite_8.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Andradite\\andradite_18.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Andradite\\andradite_28.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Andradite\\andradite_3.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Andradite\\andradite_9.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Aquamarine\\aquamarine_18.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Aquamarine\\aquamarine_3.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Aquamarine\\aquamarine_4.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Aquamarine\\aquamarine_5.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Aquamarine\\aquamarine_8.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Aventurine Green\\aventurine green_3.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Aventurine Green\\aventurine green_4.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Aventurine Green\\aventurine green_41.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Aventurine Green\\aventurine green_44.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Aventurine Green\\aventurine green_5.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Aventurine Yellow\\aventurine yellow_18.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Aventurine Yellow\\aventurine yellow_28.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Aventurine Yellow\\aventurine yellow_3.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Aventurine Yellow\\aventurine yellow_6.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Benitoite\\benitoite_18.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Benitoite\\benitoite_28.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Benitoite\\benitoite_3.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Benitoite\\benitoite_9.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Beryl Golden\\beryl golden_3.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Beryl Golden\\beryl golden_4.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Beryl Golden\\beryl golden_6.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Beryl Golden\\beryl golden_9.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Bixbite\\bixbite_3.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Bixbite\\bixbite_4.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Bixbite\\bixbite_6.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Bixbite\\bixbite_9.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Bloodstone\\bloodstone_18.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Bloodstone\\bloodstone_3.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Bloodstone\\bloodstone_5.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Bloodstone\\bloodstone_7.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Blue Lace Agate\\blue lace agate_3.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Blue Lace Agate\\blue lace agate_4.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Blue Lace Agate\\blue lace agate_41.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Blue Lace Agate\\blue lace agate_5.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Blue Lace Agate\\blue lace agate_7.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Carnelian\\carnelian_18.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Carnelian\\carnelian_28.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Carnelian\\carnelian_3.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Carnelian\\carnelian_7.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Cats Eye\\cats eye_18.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Cats Eye\\cats eye_28.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Cats Eye\\cats eye_3.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Cats Eye\\cats eye_9.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Chalcedony\\chalcedony_18.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Chalcedony\\chalcedony_2.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Chalcedony\\chalcedony_28.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Chalcedony\\chalcedony_3.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Chalcedony Blue\\chalcedony blue_18.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Chalcedony Blue\\chalcedony blue_2.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Chalcedony Blue\\chalcedony blue_28.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Chalcedony Blue\\chalcedony blue_3.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Chrome Diopside\\chrome diopside_18.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Chrome Diopside\\chrome diopside_28.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Chrome Diopside\\chrome diopside_3.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Chrome Diopside\\chrome diopside_8.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Chrysoberyl\\chrysoberyl_18.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Chrysoberyl\\chrysoberyl_28.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Chrysoberyl\\chrysoberyl_3.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Chrysoberyl\\chrysoberyl_9.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Chrysocolla\\chrysocolla_3.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Chrysocolla\\chrysocolla_4.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Chrysocolla\\chrysocolla_6.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Chrysocolla\\chrysocolla_9.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Chrysoprase\\chrysoprase_18.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Chrysoprase\\chrysoprase_2.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Chrysoprase\\chrysoprase_28.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Chrysoprase\\chrysoprase_3.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Citrine\\citrine_3.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Citrine\\citrine_4.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Citrine\\citrine_41.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Citrine\\citrine_42.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Citrine\\citrine_44.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Coral\\coral_18.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Coral\\coral_28.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Coral\\coral_3.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Coral\\coral_9.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Danburite\\danburite_18.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Danburite\\danburite_28.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Danburite\\danburite_3.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Danburite\\danburite_8.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Diamond\\diamond_18.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Diamond\\diamond_28.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Diamond\\diamond_3.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Diamond\\diamond_9.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Diaspore\\diaspore_18.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Diaspore\\diaspore_28.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Diaspore\\diaspore_3.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Diaspore\\diaspore_9.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Dumortierite\\dumortierite_18.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Dumortierite\\dumortierite_28.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Dumortierite\\dumortierite_3.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Dumortierite\\dumortierite_8.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Emerald\\emerald_18.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Emerald\\emerald_3.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Emerald\\emerald_4.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Emerald\\emerald_5.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Emerald\\emerald_8.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Fluorite\\fluorite_18.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Fluorite\\fluorite_28.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Fluorite\\fluorite_3.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Fluorite\\fluorite_8.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Garnet Red\\garnet red_3.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Garnet Red\\garnet red_4.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Garnet Red\\garnet red_6.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Garnet Red\\garnet red_9.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Goshenite\\goshenite_18.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Goshenite\\goshenite_28.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Goshenite\\goshenite_3.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Goshenite\\goshenite_9.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Grossular\\grossular_18.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Grossular\\grossular_3.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Grossular\\grossular_30.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Grossular\\grossular_32.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Hessonite\\hessonite_18.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Hessonite\\hessonite_2.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Hessonite\\hessonite_28.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Hessonite\\hessonite_3.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Hiddenite\\hiddenite_18.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Hiddenite\\hiddenite_3.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Hiddenite\\hiddenite_30.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Hiddenite\\hiddenite_32.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Iolite\\iolite_18.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Iolite\\iolite_28.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Iolite\\iolite_3.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Iolite\\iolite_8.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Jade\\jade_25.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Jade\\jade_3.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Jade\\jade_8.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Jade\\jade_9.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Jasper\\jasper_18.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Jasper\\jasper_3.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Jasper\\jasper_4.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Jasper\\jasper_5.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Jasper\\jasper_8.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Kunzite\\kunzite_18.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Kunzite\\kunzite_28.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Kunzite\\kunzite_3.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Kunzite\\kunzite_8.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Kyanite\\kyanite_18.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Kyanite\\kyanite_3.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Kyanite\\kyanite_30.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Kyanite\\kyanite_32.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Labradorite\\labradorite_3.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Labradorite\\labradorite_4.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Labradorite\\labradorite_41.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Labradorite\\labradorite_42.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Labradorite\\labradorite_44.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Lapis Lazuli\\lapis lazuli_18.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Lapis Lazuli\\lapis lazuli_28.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Lapis Lazuli\\lapis lazuli_3.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Lapis Lazuli\\lapis lazuli_9.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Larimar\\larimar_24.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Larimar\\larimar_26.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Larimar\\larimar_29.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Larimar\\larimar_8.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Malachite\\malachite_25.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Malachite\\malachite_3.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Malachite\\malachite_8.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Malachite\\malachite_9.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Moonstone\\moonstone_18.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Moonstone\\moonstone_3.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Moonstone\\moonstone_4.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Moonstone\\moonstone_5.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Moonstone\\moonstone_8.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Morganite\\morganite_18.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Morganite\\morganite_28.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Morganite\\morganite_3.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Morganite\\morganite_9.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Onyx Black\\onyx black_25.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Onyx Black\\onyx black_3.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Onyx Black\\onyx black_8.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Onyx Black\\onyx black_9.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Onyx Green\\onyx green_25.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Onyx Green\\onyx green_3.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Onyx Green\\onyx green_8.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Onyx Green\\onyx green_9.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Onyx Red\\onyx red_25.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Onyx Red\\onyx red_3.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Onyx Red\\onyx red_8.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Onyx Red\\onyx red_9.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Opal\\opal_18.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Opal\\opal_3.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Opal\\opal_4.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Opal\\opal_5.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Opal\\opal_8.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Pearl\\pearl_18.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Pearl\\pearl_28.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Pearl\\pearl_3.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Pearl\\pearl_7.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Peridot\\peridot_3.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Peridot\\peridot_4.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Peridot\\peridot_41.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Peridot\\peridot_42.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Peridot\\peridot_5.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Prehnite\\prehnite_3.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Prehnite\\prehnite_4.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Prehnite\\prehnite_41.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Prehnite\\prehnite_5.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Prehnite\\prehnite_7.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Pyrite\\pyrite_3.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Pyrite\\pyrite_4.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Pyrite\\pyrite_41.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Pyrite\\pyrite_44.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Pyrite\\pyrite_7.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Pyrope\\pyrope_18.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Pyrope\\pyrope_28.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Pyrope\\pyrope_3.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Pyrope\\pyrope_9.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Quartz Beer\\quartz beer_18.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Quartz Beer\\quartz beer_3.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Quartz Beer\\quartz beer_5.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Quartz Beer\\quartz beer_7.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Quartz Lemon\\quartz lemon_18.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Quartz Lemon\\quartz lemon_28.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Quartz Lemon\\quartz lemon_3.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Quartz Lemon\\quartz lemon_7.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Quartz Rose\\quartz rose_18.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Quartz Rose\\quartz rose_3.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Quartz Rose\\quartz rose_5.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Quartz Rose\\quartz rose_7.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Quartz Rutilated\\quartz rutilated_3.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Quartz Rutilated\\quartz rutilated_4.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Quartz Rutilated\\quartz rutilated_41.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Quartz Rutilated\\quartz rutilated_42.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Quartz Rutilated\\quartz rutilated_6.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Quartz Smoky\\quartz smoky_18.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Quartz Smoky\\quartz smoky_2.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Quartz Smoky\\quartz smoky_28.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Quartz Smoky\\quartz smoky_3.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Rhodochrosite\\rhodochrosite_18.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Rhodochrosite\\rhodochrosite_3.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Rhodochrosite\\rhodochrosite_30.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Rhodochrosite\\rhodochrosite_32.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Rhodolite\\rhodolite_3.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Rhodolite\\rhodolite_4.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Rhodolite\\rhodolite_6.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Rhodolite\\rhodolite_9.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Rhodonite\\rhodonite_18.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Rhodonite\\rhodonite_28.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Rhodonite\\rhodonite_3.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Rhodonite\\rhodonite_8.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Ruby\\ruby_18.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Ruby\\ruby_3.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Ruby\\ruby_5.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Ruby\\ruby_7.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Sapphire Blue\\sapphire blue_18.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Sapphire Blue\\sapphire blue_28.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Sapphire Blue\\sapphire blue_3.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Sapphire Blue\\sapphire blue_6.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Sapphire Pink\\sapphire pink_18.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Sapphire Pink\\sapphire pink_2.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Sapphire Pink\\sapphire pink_28.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Sapphire Pink\\sapphire pink_3.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Sapphire Purple\\sapphire purple_18.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Sapphire Purple\\sapphire purple_28.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Sapphire Purple\\sapphire purple_3.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Sapphire Purple\\sapphire purple_9.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Sapphire Yellow\\sapphire yellow_18.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Sapphire Yellow\\sapphire yellow_28.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Sapphire Yellow\\sapphire yellow_3.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Sapphire Yellow\\sapphire yellow_9.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Scapolite\\scapolite_25.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Scapolite\\scapolite_3.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Scapolite\\scapolite_8.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Scapolite\\scapolite_9.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Serpentine\\serpentine_18.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Serpentine\\serpentine_2.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Serpentine\\serpentine_28.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Serpentine\\serpentine_3.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Sodalite\\sodalite_18.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Sodalite\\sodalite_3.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Sodalite\\sodalite_5.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Sodalite\\sodalite_7.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Spessartite\\spessartite_18.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Spessartite\\spessartite_2.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Spessartite\\spessartite_28.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Spessartite\\spessartite_3.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Sphene\\sphene_18.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Sphene\\sphene_2.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Sphene\\sphene_28.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Sphene\\sphene_3.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Spinel\\spinel_25.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Spinel\\spinel_3.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Spinel\\spinel_8.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Spinel\\spinel_9.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Spodumene\\spodumene_18.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Spodumene\\spodumene_3.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Spodumene\\spodumene_5.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Spodumene\\spodumene_7.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Sunstone\\sunstone_18.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Sunstone\\sunstone_3.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Sunstone\\sunstone_30.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Sunstone\\sunstone_32.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Tanzanite\\tanzanite_3.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Tanzanite\\tanzanite_4.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Tanzanite\\tanzanite_6.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Tanzanite\\tanzanite_9.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Tigers Eye\\tigers eye_3.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Tigers Eye\\tigers eye_4.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Tigers Eye\\tigers eye_41.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Tigers Eye\\tigers eye_44.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Tigers Eye\\tigers eye_7.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Topaz\\topaz_18.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Topaz\\topaz_28.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Topaz\\topaz_3.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Topaz\\topaz_6.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Tourmaline\\tourmaline_3.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Tourmaline\\tourmaline_4.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Tourmaline\\tourmaline_41.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Tourmaline\\tourmaline_42.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Tourmaline\\tourmaline_6.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Tsavorite\\tsavorite_18.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Tsavorite\\tsavorite_28.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Tsavorite\\tsavorite_3.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Tsavorite\\tsavorite_9.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Turquoise\\turquoise_18.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Turquoise\\turquoise_28.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Turquoise\\turquoise_3.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Turquoise\\turquoise_8.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Variscite\\variscite_18.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Variscite\\variscite_2.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Variscite\\variscite_28.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Variscite\\variscite_3.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Zircon\\zircon_18.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Zircon\\zircon_28.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Zircon\\zircon_3.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Zircon\\zircon_7.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Zoisite\\zoisite_25.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Zoisite\\zoisite_3.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Zoisite\\zoisite_8.jpg\n",
      "Keeping E:/datasets/datasets_gemstone/test/Zoisite\\zoisite_9.jpg\n"
     ]
    }
   ],
   "source": [
    "import os\n",
    "from PIL import Image\n",
    "\n",
    "# Görüntü dosyalarının bulunduğu dizin\n",
    "image_dir ='E:/datasets/datasets_gemstone/test/' # Kendi dizin yolunuzu buraya ekleyin\n",
    "\n",
    "# Desteklenen görüntü formatları\n",
    "supported_formats = ('.jpg', '.jpeg', '.png')\n",
    "\n",
    "# Klasörü tarayıp görüntüleri kontrol et\n",
    "for class_dir in os.listdir(image_dir):\n",
    "    class_path = os.path.join(image_dir, class_dir)\n",
    "    \n",
    "    # Sınıf klasörü olup olmadığını kontrol et\n",
    "    if os.path.isdir(class_path):\n",
    "        for img_name in os.listdir(class_path):\n",
    "            img_path = os.path.join(class_path, img_name)\n",
    "            \n",
    "            # Dosya formatını kontrol et\n",
    "            if img_name.lower().endswith(supported_formats):\n",
    "                try:\n",
    "                    # Görüntüyü aç\n",
    "                    with Image.open(img_path) as img:\n",
    "                        # Görüntü boyutlarını kontrol et\n",
    "                        if img.mode != 'RGB':\n",
    "                            print(f\"Deleting {img_path} (not RGB)\")\n",
    "                            os.remove(img_path)  # 3 kanallı olmayan görüntüyü sil\n",
    "                        else:\n",
    "                            print(f\"Keeping {img_path}\")  # 3 kanallı görüntüyü tut\n",
    "                except Exception as e:\n",
    "                    print(f\"Error processing {img_path}: {e}\")\n",
    "                    # Hata durumunda dosyayı silebilirsiniz (isteğe bağlı)\n",
    "                    # os.remove(img_path)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 105,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "                                               image        label\n",
      "0  <PIL.Image.Image image mode=RGB size=225x225 a...  Alexandrite\n",
      "1  <PIL.Image.Image image mode=RGB size=251x228 a...  Alexandrite\n",
      "2  <PIL.Image.Image image mode=RGB size=347x347 a...  Alexandrite\n",
      "3  <PIL.Image.Image image mode=RGB size=218x231 a...  Alexandrite\n",
      "4  <PIL.Image.Image image mode=RGB size=395x395 a...  Alexandrite\n"
     ]
    }
   ],
   "source": [
    "import os\n",
    "import pandas as pd\n",
    "from PIL import Image\n",
    "\n",
    "# Dataset yolunu belirtin\n",
    "dataset_dir = 'E:/datasets/datasets_gemstone/train/'\n",
    "\n",
    "train_data = []\n",
    "train_labels = []\n",
    "\n",
    "# Eğitim verilerini yükleme\n",
    "for class_name in os.listdir(dataset_dir):\n",
    "    class_dir = os.path.join(dataset_dir, class_name)\n",
    "    if os.path.isdir(class_dir):  # Sadece dizinleri kontrol et\n",
    "        for img_name in os.listdir(class_dir):\n",
    "            img_path = os.path.join(class_dir, img_name)\n",
    "            # Görüntüyü yükle\n",
    "            with Image.open(img_path) as image:\n",
    "                train_data.append(image.copy())  # Görüntüyü kopyala\n",
    "                train_labels.append(class_name)  # Sınıf adını etiket olarak ekle\n",
    "\n",
    "# DataFrame oluşturma\n",
    "df = pd.DataFrame({\n",
    "    'image': train_data,\n",
    "    'label': train_labels\n",
    "})\n",
    "\n",
    "# DataFrame'i görüntüleme\n",
    "print(df.head())\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 106,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_train=df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 107,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "                                               image        label\n",
      "0  <PIL.Image.Image image mode=RGB size=225x225 a...  Alexandrite\n",
      "1  <PIL.Image.Image image mode=RGB size=251x228 a...  Alexandrite\n",
      "2  <PIL.Image.Image image mode=RGB size=347x347 a...  Alexandrite\n",
      "3  <PIL.Image.Image image mode=RGB size=218x231 a...  Alexandrite\n",
      "4  <PIL.Image.Image image mode=RGB size=395x395 a...  Alexandrite\n"
     ]
    }
   ],
   "source": [
    "import os\n",
    "import pandas as pd\n",
    "from PIL import Image\n",
    "\n",
    "# Dataset yolunu belirtin\n",
    "dataset_dir = 'E:/datasets/datasets_gemstone/test/'\n",
    "\n",
    "test_data = []\n",
    "test_labels = []\n",
    "\n",
    "# Eğitim verilerini yükleme\n",
    "for class_name in os.listdir(dataset_dir):\n",
    "    class_dir = os.path.join(dataset_dir, class_name)\n",
    "    if os.path.isdir(class_dir):  # Sadece dizinleri kontrol et\n",
    "        for img_name in os.listdir(class_dir):\n",
    "            img_path = os.path.join(class_dir, img_name)\n",
    "            # Görüntüyü yükle\n",
    "            with Image.open(img_path) as image:\n",
    "                test_data.append(image.copy())  # Görüntüyü kopyala\n",
    "                test_labels.append(class_name)  # Sınıf adını etiket olarak ekle\n",
    "\n",
    "# DataFrame oluşturma\n",
    "df_test = pd.DataFrame({\n",
    "    'image': test_data,\n",
    "    'label': test_labels\n",
    "})\n",
    "\n",
    "# DataFrame'i görüntüleme\n",
    "print(df.head())\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 108,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "                                                  image        label\n",
      "0     <PIL.Image.Image image mode=RGB size=225x225 a...  Alexandrite\n",
      "1     <PIL.Image.Image image mode=RGB size=251x228 a...  Alexandrite\n",
      "2     <PIL.Image.Image image mode=RGB size=347x347 a...  Alexandrite\n",
      "3     <PIL.Image.Image image mode=RGB size=218x231 a...  Alexandrite\n",
      "4     <PIL.Image.Image image mode=RGB size=395x395 a...  Alexandrite\n",
      "...                                                 ...          ...\n",
      "3211  <PIL.Image.Image image mode=RGB size=314x410 a...       Zircon\n",
      "3212  <PIL.Image.Image image mode=RGB size=225x225 a...      Zoisite\n",
      "3213  <PIL.Image.Image image mode=RGB size=420x420 a...      Zoisite\n",
      "3214  <PIL.Image.Image image mode=RGB size=225x225 a...      Zoisite\n",
      "3215  <PIL.Image.Image image mode=RGB size=454x454 a...      Zoisite\n",
      "\n",
      "[3216 rows x 2 columns]\n"
     ]
    }
   ],
   "source": [
    "combined_df = pd.concat([df_train,df_test], ignore_index=True)\n",
    "\n",
    "print(combined_df)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 109,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>image</th>\n",
       "      <th>label</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>&lt;PIL.Image.Image image mode=RGB size=225x225 a...</td>\n",
       "      <td>Alexandrite</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>&lt;PIL.Image.Image image mode=RGB size=251x228 a...</td>\n",
       "      <td>Alexandrite</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>&lt;PIL.Image.Image image mode=RGB size=347x347 a...</td>\n",
       "      <td>Alexandrite</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>&lt;PIL.Image.Image image mode=RGB size=218x231 a...</td>\n",
       "      <td>Alexandrite</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>&lt;PIL.Image.Image image mode=RGB size=395x395 a...</td>\n",
       "      <td>Alexandrite</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3211</th>\n",
       "      <td>&lt;PIL.Image.Image image mode=RGB size=314x410 a...</td>\n",
       "      <td>Zircon</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3212</th>\n",
       "      <td>&lt;PIL.Image.Image image mode=RGB size=225x225 a...</td>\n",
       "      <td>Zoisite</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3213</th>\n",
       "      <td>&lt;PIL.Image.Image image mode=RGB size=420x420 a...</td>\n",
       "      <td>Zoisite</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3214</th>\n",
       "      <td>&lt;PIL.Image.Image image mode=RGB size=225x225 a...</td>\n",
       "      <td>Zoisite</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3215</th>\n",
       "      <td>&lt;PIL.Image.Image image mode=RGB size=454x454 a...</td>\n",
       "      <td>Zoisite</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>3216 rows × 2 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                  image        label\n",
       "0     <PIL.Image.Image image mode=RGB size=225x225 a...  Alexandrite\n",
       "1     <PIL.Image.Image image mode=RGB size=251x228 a...  Alexandrite\n",
       "2     <PIL.Image.Image image mode=RGB size=347x347 a...  Alexandrite\n",
       "3     <PIL.Image.Image image mode=RGB size=218x231 a...  Alexandrite\n",
       "4     <PIL.Image.Image image mode=RGB size=395x395 a...  Alexandrite\n",
       "...                                                 ...          ...\n",
       "3211  <PIL.Image.Image image mode=RGB size=314x410 a...       Zircon\n",
       "3212  <PIL.Image.Image image mode=RGB size=225x225 a...      Zoisite\n",
       "3213  <PIL.Image.Image image mode=RGB size=420x420 a...      Zoisite\n",
       "3214  <PIL.Image.Image image mode=RGB size=225x225 a...      Zoisite\n",
       "3215  <PIL.Image.Image image mode=RGB size=454x454 a...      Zoisite\n",
       "\n",
       "[3216 rows x 2 columns]"
      ]
     },
     "execution_count": 109,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "combined_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 110,
   "metadata": {},
   "outputs": [],
   "source": [
    "from torch.utils.data import Dataset ,DataLoader\n",
    "from torchvision import transforms\n",
    "from sklearn.preprocessing import LabelEncoder"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 111,
   "metadata": {},
   "outputs": [],
   "source": [
    "label_encoder=LabelEncoder()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 112,
   "metadata": {},
   "outputs": [],
   "source": [
    "combined_df[\"label\"]=label_encoder.fit_transform(combined_df[\"label\"])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 113,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>image</th>\n",
       "      <th>label</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>&lt;PIL.Image.Image image mode=RGB size=225x225 a...</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>&lt;PIL.Image.Image image mode=RGB size=251x228 a...</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>&lt;PIL.Image.Image image mode=RGB size=347x347 a...</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>&lt;PIL.Image.Image image mode=RGB size=218x231 a...</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>&lt;PIL.Image.Image image mode=RGB size=395x395 a...</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3211</th>\n",
       "      <td>&lt;PIL.Image.Image image mode=RGB size=314x410 a...</td>\n",
       "      <td>85</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3212</th>\n",
       "      <td>&lt;PIL.Image.Image image mode=RGB size=225x225 a...</td>\n",
       "      <td>86</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3213</th>\n",
       "      <td>&lt;PIL.Image.Image image mode=RGB size=420x420 a...</td>\n",
       "      <td>86</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3214</th>\n",
       "      <td>&lt;PIL.Image.Image image mode=RGB size=225x225 a...</td>\n",
       "      <td>86</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3215</th>\n",
       "      <td>&lt;PIL.Image.Image image mode=RGB size=454x454 a...</td>\n",
       "      <td>86</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>3216 rows × 2 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                  image  label\n",
       "0     <PIL.Image.Image image mode=RGB size=225x225 a...      0\n",
       "1     <PIL.Image.Image image mode=RGB size=251x228 a...      0\n",
       "2     <PIL.Image.Image image mode=RGB size=347x347 a...      0\n",
       "3     <PIL.Image.Image image mode=RGB size=218x231 a...      0\n",
       "4     <PIL.Image.Image image mode=RGB size=395x395 a...      0\n",
       "...                                                 ...    ...\n",
       "3211  <PIL.Image.Image image mode=RGB size=314x410 a...     85\n",
       "3212  <PIL.Image.Image image mode=RGB size=225x225 a...     86\n",
       "3213  <PIL.Image.Image image mode=RGB size=420x420 a...     86\n",
       "3214  <PIL.Image.Image image mode=RGB size=225x225 a...     86\n",
       "3215  <PIL.Image.Image image mode=RGB size=454x454 a...     86\n",
       "\n",
       "[3216 rows x 2 columns]"
      ]
     },
     "execution_count": 113,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "combined_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 114,
   "metadata": {},
   "outputs": [],
   "source": [
    "num_classes=combined_df['label'].nunique()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 115,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "87"
      ]
     },
     "execution_count": 115,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "num_classes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 116,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(87,)\n"
     ]
    }
   ],
   "source": [
    "print(label_encoder.classes_.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 117,
   "metadata": {},
   "outputs": [],
   "source": [
    "import torch \n",
    "from torch.utils.data import Dataset\n",
    "from PIL import Image"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 118,
   "metadata": {},
   "outputs": [],
   "source": [
    "class CustomDataset(Dataset):\n",
    "    def __init__ (self,dataframe,transform=None):\n",
    "        self.data=dataframe\n",
    "        self.transform=transform\n",
    "    def __len__(self):\n",
    "\n",
    "        return len(self.data)\n",
    "\n",
    "        \n",
    "    def __getitem__(self,index):\n",
    "\n",
    "        label=self.data.iloc[index][\"label\"]\n",
    "\n",
    "        image=self.data.iloc[index][\"image\"]\n",
    "\n",
    "        if self.transform:\n",
    "            image=self.transform(image)\n",
    "\n",
    "\n",
    "        return image, label\n",
    "\n",
    "\n",
    "\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 119,
   "metadata": {},
   "outputs": [],
   "source": [
    "from torchvision import transforms"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 120,
   "metadata": {},
   "outputs": [],
   "source": [
    "dataframe=combined_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 121,
   "metadata": {},
   "outputs": [],
   "source": [
    "transform = transforms.Compose([\n",
    "    \n",
    "    transforms.Resize((224, 224)),  # RESNET İÇİN\n",
    "    transforms.ToTensor(),  # Görüntüyü tensöre çeviriyoruz\n",
    "    transforms.Normalize(mean=[0.485, 0.456, 0.406],  # Normalizasyon\n",
    "                         std=[0.229, 0.224, 0.225])\n",
    "])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 122,
   "metadata": {},
   "outputs": [],
   "source": [
    "dataset=CustomDataset(dataframe=dataframe,transform=transform)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 123,
   "metadata": {},
   "outputs": [],
   "source": [
    "image,label=dataset[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 124,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "torch.Size([3, 224, 224])"
      ]
     },
     "execution_count": 124,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "image.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 125,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<__main__.CustomDataset at 0x237127c83a0>"
      ]
     },
     "execution_count": 125,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 126,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.model_selection import train_test_split"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 127,
   "metadata": {},
   "outputs": [],
   "source": [
    "labels = dataset.data['label'].values[:len(dataset)]  \n",
    "\n",
    "indices = list(range(len(dataset)))\n",
    "train_indices, test_indices = train_test_split(indices, test_size=0.2, stratify=labels)\n",
    "\n",
    "train_dataset = torch.utils.data.Subset(dataset, train_indices)\n",
    "test_dataset = torch.utils.data.Subset(dataset, test_indices)\n",
    "\n",
    "train_loader = DataLoader(train_dataset, batch_size=64, shuffle=True)\n",
    "test_loader = DataLoader(test_dataset, batch_size=64, shuffle=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 128,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "tensor([78, 83, 60, 74, 62, 49,  4, 85, 73, 68, 33, 65, 76, 55, 60, 78,  0, 18,\n",
      "        46,  0,  0, 39, 12, 57, 39,  3, 57, 55, 10, 38, 68, 75, 69, 55, 34, 58,\n",
      "        21, 60, 27, 60, 52,  6, 79, 10, 10, 19, 28, 60, 57, 68, 40, 83, 10, 71,\n",
      "        69, 23, 73, 12, 57, 76, 23, 26,  4, 25], dtype=torch.int32)\n",
      "tensor([80, 54, 54, 48, 43, 36, 55,  3, 39, 44,  9, 79, 33, 74, 81, 29, 24,  8,\n",
      "         5, 73, 48, 61, 72, 43, 38, 35, 82, 42, 75, 67, 25, 80, 24, 19, 56, 53,\n",
      "        45, 19, 32, 36, 58, 76, 81, 14, 28, 52, 66, 19, 17, 77, 69, 28, 78, 17,\n",
      "        19, 15,  3, 37, 72, 44, 66, 76, 52, 22], dtype=torch.int32)\n",
      "tensor([86,  2, 79, 19, 24, 14,  4, 12,  2,  8, 21, 19, 50, 79, 84, 24, 24, 72,\n",
      "        29,  2, 53, 55, 63, 14, 81, 52, 15, 79, 53, 11, 40, 83, 30, 46, 30, 13,\n",
      "        11, 14,  7, 81, 69, 13, 27, 74, 10, 43, 14, 42,  9, 64, 38, 26, 51, 66,\n",
      "        34, 77, 67,  5, 79,  7, 26, 60,  6, 13], dtype=torch.int32)\n",
      "tensor([18, 84, 41, 45, 36, 45, 63, 82, 37, 31, 31, 14, 24, 71,  7, 15, 83, 38,\n",
      "        44, 82,  1, 37, 63, 82, 58, 28, 38, 41,  0, 82, 60, 56, 69, 74,  6,  3,\n",
      "        53, 56, 54, 18, 19, 81, 43, 57, 54, 10, 15, 57, 77, 63, 52, 50, 33, 47,\n",
      "         6, 30, 56, 47, 45, 13, 41, 62, 36, 30], dtype=torch.int32)\n",
      "tensor([52, 31, 27, 77, 25, 20, 30, 50, 15, 29, 36, 28, 29, 32, 39, 12, 80,  8,\n",
      "        60, 79, 22, 75, 30, 32, 64, 31,  9,  5, 48, 40, 84, 70, 12, 77, 74, 34,\n",
      "        85,  9, 25, 60, 81, 52, 12, 76, 82, 63, 32, 46, 41, 70, 12, 15, 43, 24,\n",
      "        54, 33, 37, 23, 43, 75, 16, 72, 14, 32], dtype=torch.int32)\n",
      "tensor([ 6, 44, 63, 41, 16, 55, 32,  7, 28,  7, 28, 65, 34, 20, 50, 58, 11, 74,\n",
      "        20, 65,  5, 75, 71, 47, 31,  5, 49,  6, 50, 47, 74, 11, 42, 84,  1, 54,\n",
      "        55, 68,  5, 79, 65, 12, 11, 48, 31, 51, 70, 22, 42, 76, 19,  0, 15, 57,\n",
      "        36, 85,  9, 49, 82, 15, 64, 13, 65, 64], dtype=torch.int32)\n",
      "tensor([40, 42, 61, 52, 44, 77, 69, 86, 83,  0, 80, 63, 43, 60, 42, 68, 73, 32,\n",
      "        67, 72, 62,  6, 44, 67, 84, 33, 20, 60, 34, 41, 34, 14, 42, 58, 17, 85,\n",
      "        11, 36, 19, 44, 23, 21,  1, 58, 37, 35, 77, 72,  6, 53, 20, 77, 66, 18,\n",
      "        76, 50, 77, 26, 16, 23, 13,  8, 14, 43], dtype=torch.int32)\n",
      "tensor([20, 85,  5, 10, 49, 41,  7, 60,  3,  9, 86, 57, 74, 67, 49, 59, 66, 21,\n",
      "        61, 79, 21, 37, 64, 15,  5, 46, 60, 50, 30, 32, 70,  4, 66, 85,  1, 62,\n",
      "        44, 43, 23, 26, 59, 40, 40, 22, 15, 19, 38, 49, 36, 53, 16, 51, 77, 72,\n",
      "        11, 28,  5, 10,  0,  1, 51, 31, 54,  5], dtype=torch.int32)\n",
      "tensor([13, 13, 78, 64, 53,  5, 65,  6, 54, 13,  2, 75, 82, 72, 63,  7,  4, 69,\n",
      "        42, 67, 41, 38, 48, 54, 66, 71, 28, 14, 55, 21, 78, 10, 32,  4,  0, 20,\n",
      "        13,  6, 21, 72, 86, 32, 68, 65, 24, 54,  9, 57,  3, 51, 45, 50, 33, 46,\n",
      "        73, 63, 45, 58, 83, 81, 79, 35, 45, 36], dtype=torch.int32)\n",
      "tensor([66, 16, 43, 10, 23, 45, 50, 10, 39, 43,  0, 36, 70, 30, 24, 49,  6, 17,\n",
      "        11, 83, 74, 12, 43, 57, 30, 27, 38, 35, 24, 55, 39, 40,  6, 34, 42, 51,\n",
      "         8, 34,  2, 27, 52, 85,  3, 32, 50, 38, 18, 61, 47, 43, 83,  8, 84, 70,\n",
      "        59,  6, 59, 67, 37, 66, 81, 62, 30, 62], dtype=torch.int32)\n",
      "tensor([33, 80, 15, 46, 68, 42,  0, 75, 67, 16,  2, 83, 59, 18, 73, 11, 26, 60,\n",
      "        42,  1, 22,  2, 25, 29, 74, 78, 32, 41, 15,  2, 22,  2, 78, 58, 31, 30,\n",
      "        84, 51, 65, 80, 45, 72, 73, 80, 59, 60, 76, 64, 83, 39, 69, 77, 52, 67,\n",
      "        74, 42, 54, 63, 27, 14, 24, 35, 24, 74], dtype=torch.int32)\n",
      "tensor([58, 46, 52, 67, 20, 25, 60,  6, 47, 33, 61, 54, 31, 42, 58, 81, 78, 37,\n",
      "        22, 25, 61, 69, 43, 36, 55, 54, 81, 64, 26, 78,  8, 58,  1, 80, 70,  3,\n",
      "        66, 39,  9, 13, 83, 72, 20, 60, 53, 69, 29, 35,  8, 67,  6,  4, 56, 69,\n",
      "        47, 40, 29, 32, 34, 57, 19, 47, 16, 21], dtype=torch.int32)\n",
      "tensor([11, 40, 34, 71, 44, 86, 36, 79, 56, 80, 28, 84, 55, 62, 42, 42, 80,  1,\n",
      "        54,  8, 63,  3, 12,  9, 16, 12, 62, 44, 37, 43, 23, 61, 18,  3,  1, 31,\n",
      "        67, 23, 17, 78, 82, 42, 44, 57, 56, 68,  6, 38, 53, 15, 60, 22,  0, 67,\n",
      "        58, 30, 62, 15,  8, 51, 24, 67, 47, 68], dtype=torch.int32)\n",
      "tensor([31, 58, 48,  1, 28, 12, 55, 29, 41, 21, 76,  0, 32, 39, 80, 58, 20, 26,\n",
      "        22, 37, 86,  2, 62, 59,  7, 57, 69, 47, 31, 17, 54, 86,  9, 73, 81, 47,\n",
      "        74, 17, 17, 66, 85, 26, 39, 14, 75, 58, 73, 25, 10,  8, 54, 61,  9, 83,\n",
      "        30, 29, 11, 13, 48, 14, 65, 39, 57, 83], dtype=torch.int32)\n",
      "tensor([79,  1, 30,  0, 81, 29, 22, 36, 79, 62, 79, 26, 42, 14, 27, 11, 41, 41,\n",
      "        50, 24, 15, 23, 20, 70, 44, 51, 29, 72, 65, 81, 68, 84, 78, 77, 45, 22,\n",
      "        47, 86, 33, 62, 70, 60, 45, 16, 23, 37, 84, 55, 31, 80, 38, 50,  5, 32,\n",
      "        23, 28, 50, 28, 56, 33, 17, 70, 47, 73], dtype=torch.int32)\n",
      "tensor([71, 19, 51, 50, 30, 27, 46, 11, 10, 84, 65, 42, 41, 86, 39, 75,  7, 10,\n",
      "        71,  9, 82, 37, 39, 39, 80, 18, 51, 38, 71, 12, 63, 31, 26, 66, 20, 39,\n",
      "        35, 53, 51, 26, 29, 76, 46,  4, 13, 11, 24, 22,  5, 80, 41, 69, 37, 43,\n",
      "        31, 43, 76, 44, 21, 17, 27, 63, 58, 34], dtype=torch.int32)\n",
      "tensor([ 9, 10, 64,  9, 42, 79, 29,  9,  9, 61, 85, 62, 56, 76, 49, 58,  8, 49,\n",
      "        22,  7, 15, 50, 49, 64,  4, 82, 11, 25, 81, 47, 39,  1, 36, 12, 53, 86,\n",
      "        64, 56, 85, 36, 38, 75, 35, 20, 60, 18, 64, 28,  0,  0,  0, 25, 82, 73,\n",
      "        45, 79, 42, 32, 62, 80, 42, 62, 15, 76], dtype=torch.int32)\n",
      "tensor([21, 53, 13,  3, 49, 13, 70, 54, 35,  8, 72, 70, 51, 22,  3, 17, 82, 69,\n",
      "         2,  2, 66, 19, 79, 63, 73, 77, 61,  9, 75, 49, 51, 31, 16, 31,  0, 78,\n",
      "        71, 48, 52,  1, 37, 68, 79, 62, 67, 79, 11, 22, 71,  0, 42, 37, 83, 71,\n",
      "        31, 77, 77, 11, 22, 12, 46, 44, 67, 39], dtype=torch.int32)\n",
      "tensor([22, 45, 57, 60, 75, 34, 32, 81, 82, 37,  3,  9, 33, 83, 59, 77, 72,  4,\n",
      "        67, 84, 24, 55, 11, 59,  0, 80, 82, 49, 38, 24, 59,  4, 15, 41, 70, 37,\n",
      "         9, 76, 22, 27, 73, 32,  3, 24, 72, 33, 84, 28, 53, 51, 21, 40, 21, 13,\n",
      "         7, 13, 56, 66, 63, 62, 40, 29, 46, 53], dtype=torch.int32)\n",
      "tensor([52, 32, 34, 59, 31,  3,  3, 40, 83, 29, 47, 74, 85,  0,  6, 72, 46, 39,\n",
      "        24,  5, 25, 37,  2, 22, 20,  9, 45,  7, 83,  9,  8, 13, 27,  9, 55, 14,\n",
      "        20, 84, 56, 64, 53, 79, 25, 30, 55, 74, 63, 48,  6, 78, 54, 51, 37, 68,\n",
      "        48, 37, 64, 71, 76, 16, 50,  9, 22, 26], dtype=torch.int32)\n",
      "tensor([81, 23, 85, 55, 17,  3, 71,  4, 27, 86, 55,  0, 16,  8, 72, 33, 29, 59,\n",
      "        16,  6, 56, 44, 49, 31, 51, 44, 15, 70, 13,  8, 45, 77, 68, 21, 35, 56,\n",
      "        54, 44, 32, 61, 51, 63, 30, 54, 70, 17, 76, 74, 24,  5,  1, 11, 79, 33,\n",
      "        14, 75, 54, 48, 79, 79, 28, 49, 20, 27], dtype=torch.int32)\n",
      "tensor([30, 13, 59, 13, 63, 57, 78, 27, 70, 70,  1, 37, 80, 32, 15,  8,  7, 61,\n",
      "        81, 57, 46, 51, 63, 81, 73, 24, 78, 52, 30, 65, 13, 47, 29, 22, 42, 23,\n",
      "        59, 37, 67, 81, 67, 65, 49, 29, 53, 35, 64, 68, 76, 60, 16,  2, 83, 12,\n",
      "        82,  1, 50, 40, 52, 82, 82, 39, 50, 18], dtype=torch.int32)\n",
      "tensor([12, 67,  9, 39, 49,  9, 51, 53,  7, 48, 54,  9, 48, 11, 16, 43, 65, 32,\n",
      "        17, 61, 76, 54, 81, 18, 63, 13, 83, 34,  0, 53,  8, 10, 49, 47, 13, 62,\n",
      "         4, 45, 46, 51, 80, 60, 34, 72, 61, 86, 57, 46, 86, 55, 55, 68, 17, 82,\n",
      "        61, 10, 26, 27, 30,  0, 69, 57, 17, 16], dtype=torch.int32)\n",
      "tensor([ 7, 71, 21, 71, 81, 79, 85, 81,  3, 86, 58, 79,  4,  2, 68, 35, 76,  5,\n",
      "        69,  0, 40, 18, 45, 36, 33, 80, 83, 15,  3, 29, 85, 17, 14, 38, 61, 44,\n",
      "        11,  7, 58, 15, 60, 46, 66, 76, 76, 11, 47, 55, 66, 79, 21, 77, 85, 62,\n",
      "        72, 77, 16,  9, 73, 24, 81, 14, 69, 39], dtype=torch.int32)\n",
      "tensor([10, 13, 85, 22, 65, 64, 36, 39, 57, 39,  8, 30, 28, 20, 84, 16, 82, 44,\n",
      "        51, 47, 56, 15, 61, 81, 13, 28, 23, 50, 55, 20, 72, 19, 27, 51, 86, 48,\n",
      "        50, 53, 74, 53, 73, 29, 10, 20, 32, 44, 12, 63, 67, 33, 55, 85, 12, 58,\n",
      "         9, 45, 68, 67, 66,  3, 54, 24, 74, 64], dtype=torch.int32)\n",
      "tensor([46, 76, 80, 23, 51, 49, 47, 59, 14, 20, 41, 47, 35, 38, 25, 33, 40, 62,\n",
      "        37, 59, 60, 83, 25, 37,  6, 38, 59, 53, 56, 26, 39, 26, 35, 78, 32, 41,\n",
      "        38, 46,  7, 29, 53, 58, 68, 73, 30, 48, 41,  9,  8, 44, 27, 47, 61, 55,\n",
      "        47, 43, 72,  8, 25, 12, 39, 28, 10, 21], dtype=torch.int32)\n",
      "tensor([40, 32,  1, 65, 72, 15, 68, 85, 14, 68, 70, 85, 71, 41, 74,  6, 56, 27,\n",
      "        56, 61, 13,  5, 85, 45, 72, 43, 63,  4, 72, 45, 56,  9, 74, 56, 42, 75,\n",
      "        34, 61,  9, 85, 20, 69, 65, 79, 77, 46, 55, 86,  3, 82, 64, 80,  4, 10,\n",
      "         5, 33, 52, 68, 16, 63, 41, 77, 43, 40], dtype=torch.int32)\n",
      "tensor([38, 21, 17, 40, 56, 80, 71, 73, 84, 28,  3, 56, 58, 85, 46, 34, 17,  2,\n",
      "        77, 81, 69, 82, 30, 46, 12, 85, 35,  0, 49, 17, 22, 77, 75, 43, 14, 28,\n",
      "         8, 55, 66,  5, 70, 48, 29,  5, 36,  1, 69, 85, 26, 71, 65,  3, 17, 59,\n",
      "        19, 16, 24, 34, 77,  3, 85, 46, 26, 31], dtype=torch.int32)\n",
      "tensor([53, 86, 31, 30, 26, 65, 27, 67, 30, 45, 15, 20, 28, 29, 25, 57, 14, 74,\n",
      "        68, 65, 57, 42, 51, 20, 32,  4, 55, 60, 86, 36, 29,  4, 57,  6, 75, 80,\n",
      "        49, 22, 30, 16, 68, 74, 66, 77, 86, 35, 63,  8, 69, 20, 75, 42, 32, 12,\n",
      "        78, 54, 14, 84, 48, 83,  2, 76, 35, 74], dtype=torch.int32)\n",
      "tensor([79, 76,  1, 56,  5, 33, 16,  8, 81,  1, 54, 37, 42, 63, 35, 64, 86, 52,\n",
      "        68, 28, 10, 52, 25,  3, 38, 15, 61, 13, 20,  7, 71,  6, 14, 46, 45, 69,\n",
      "        81, 79, 57, 42, 23, 59,  8,  9, 41, 78, 42, 64, 62, 11, 60, 28, 52, 53,\n",
      "        26, 58, 72, 32, 55, 58, 39, 17, 27, 63], dtype=torch.int32)\n",
      "tensor([59, 36, 83, 49, 55,  1, 24,  3, 52, 75,  7,  4, 15, 33, 13, 24, 76, 27,\n",
      "        22, 60,  5, 84, 59, 76, 14, 70, 51,  8, 56, 42, 66, 55, 22, 61, 22, 43,\n",
      "        30, 56, 85, 38, 50, 65,  2, 18, 55, 71, 12, 50, 39, 52, 66, 82, 64, 30,\n",
      "         5, 68,  1, 71, 27, 12, 35,  7, 36, 46], dtype=torch.int32)\n",
      "tensor([75, 63, 72, 40, 35, 33, 53, 41, 30, 62, 50, 73, 69, 80, 48, 10, 33, 73,\n",
      "        66, 55, 55, 32, 19,  9, 54, 37, 19, 21, 23,  5, 45, 59, 74, 70, 18, 83,\n",
      "         8, 25, 16,  2, 52, 52, 15, 80, 30, 35, 52, 54, 41,  4, 24, 36, 24, 64,\n",
      "        18, 59, 68, 33, 36, 84, 78, 51, 12,  8], dtype=torch.int32)\n",
      "tensor([ 1, 16, 44, 63, 64, 70, 20, 73, 79, 58,  2, 52, 34, 26, 86, 60,  8, 48,\n",
      "        70, 32, 46, 55, 75, 73, 30, 58, 15,  7, 66, 40, 80, 27, 37,  4, 36, 62,\n",
      "        72, 27, 71, 74,  9, 31, 84, 34, 26, 73, 65,  7, 73,  0, 65,  8, 25,  9,\n",
      "         4, 81, 55, 25, 40, 66, 17, 14,  7, 60], dtype=torch.int32)\n",
      "tensor([74, 12, 42, 85, 82, 79, 84, 53, 10, 59, 18, 43, 55,  9, 56, 19,  0, 54,\n",
      "        83, 43, 75, 29, 11, 53, 51, 59, 50, 57,  0, 59, 21, 59, 65, 52, 46,  9,\n",
      "        33, 36, 13, 83, 18, 12, 26, 76, 40, 19, 20, 78, 71, 35, 46, 44,  2,  5,\n",
      "        65, 68, 63, 85, 32, 24, 72, 43, 46,  4], dtype=torch.int32)\n",
      "tensor([44, 74, 49, 18,  6, 20, 22, 75, 69, 26, 31, 35, 51, 21,  1,  5, 39, 29,\n",
      "        48, 79, 65, 11, 80, 80, 76, 51, 18, 24, 30, 25, 85, 66, 19, 56,  7,  6,\n",
      "        52, 34,  7, 34, 10, 34, 41,  0,  2, 55, 18, 10,  5, 26, 28, 19, 60,  4,\n",
      "        73,  2,  4,  6, 40, 59, 25,  7, 14, 30], dtype=torch.int32)\n",
      "tensor([19, 18, 78, 44, 52, 60,  8, 53,  5, 83, 16,  9, 78, 12, 82, 78, 69, 71,\n",
      "        54, 45, 29, 17, 82, 79, 58, 49, 18, 54, 41, 79, 78, 66, 16, 54, 39, 69,\n",
      "        51,  8, 54, 64, 58,  2,  3, 15,  7, 23, 38, 84,  2, 72, 28,  4, 75,  2,\n",
      "        37, 85, 65, 28, 28, 12, 27, 59, 83, 84], dtype=torch.int32)\n",
      "tensor([33, 75, 24, 81, 76, 21, 67, 48, 64, 15,  4, 79, 41, 48,  2, 46,  4, 39,\n",
      "        81,  8, 81, 47, 10, 78, 81, 64, 78, 42, 14, 78, 81, 31, 84, 34,  9, 59,\n",
      "        55, 42, 24, 58, 80,  1, 52, 66, 81, 23, 25, 38, 16, 61, 31, 53,  9, 65,\n",
      "        33, 50, 11, 17, 21, 64, 33, 79, 54, 27], dtype=torch.int32)\n",
      "tensor([58, 76, 21, 22, 72, 86, 20, 11, 10, 19, 37, 77,  2, 34,  7,  1, 78, 13,\n",
      "        32, 64, 16, 57,  3,  6, 52, 15, 17, 22, 64, 60, 40, 53, 40, 71, 31, 80,\n",
      "        75, 61, 42, 48, 21,  4, 26, 50, 84, 39, 62, 42, 33, 73, 29, 10,  5, 46,\n",
      "        24, 25, 43, 63, 35, 17, 78, 84, 48, 14], dtype=torch.int32)\n",
      "tensor([78, 19, 26, 19, 12, 16, 42, 61, 53, 65, 68, 67, 79, 66, 40, 57, 53, 12,\n",
      "        70, 76, 48, 65, 71, 73, 17, 72, 51, 23, 27, 18, 14, 11, 23, 67, 23, 83,\n",
      "        57,  6, 75, 39, 45, 47, 24, 35, 53, 21, 71, 22, 57, 25, 47, 35, 53, 46,\n",
      "        18, 59, 38, 56, 40, 26, 15, 10, 18, 47], dtype=torch.int32)\n",
      "tensor([ 4, 24, 21, 60, 40, 12, 65, 82,  4, 39, 29,  1, 66, 78, 34, 49, 40, 18,\n",
      "        67, 61, 46, 49, 57, 35, 25, 57, 23, 53, 36,  0, 48, 45, 43, 22, 63, 19,\n",
      "        14, 16, 15, 21, 23, 31, 31,  5, 47, 27, 51, 26, 69, 18, 66, 38, 13, 25,\n",
      "         6, 23,  5, 38, 62, 62,  2, 82, 24, 69], dtype=torch.int32)\n",
      "tensor([61, 86, 78, 86, 70, 25, 79, 23, 86, 63, 70, 46], dtype=torch.int32)\n"
     ]
    }
   ],
   "source": [
    "for x,y in train_loader: \n",
    "    print(y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 129,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train dataset size: 2572\n",
      "Test dataset size: 644\n"
     ]
    }
   ],
   "source": [
    "print(f\"Train dataset size: {len(train_dataset)}\")\n",
    "print(f\"Test dataset size: {len(test_dataset)}\")\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 130,
   "metadata": {},
   "outputs": [],
   "source": [
    "if torch.cuda.is_available():\n",
    "    device=torch.device(\"cuda\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 131,
   "metadata": {},
   "outputs": [],
   "source": [
    "device\n",
    "import torch.nn as nn"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 132,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "c:\\Users\\ninja\\AppData\\Local\\Programs\\Python\\Python39\\lib\\site-packages\\torchvision\\models\\_utils.py:208: UserWarning: The parameter 'pretrained' is deprecated since 0.13 and may be removed in the future, please use 'weights' instead.\n",
      "  warnings.warn(\n",
      "c:\\Users\\ninja\\AppData\\Local\\Programs\\Python\\Python39\\lib\\site-packages\\torchvision\\models\\_utils.py:223: UserWarning: Arguments other than a weight enum or `None` for 'weights' are deprecated since 0.13 and may be removed in the future. The current behavior is equivalent to passing `weights=ResNet18_Weights.IMAGENET1K_V1`. You can also use `weights=ResNet18_Weights.DEFAULT` to get the most up-to-date weights.\n",
      "  warnings.warn(msg)\n"
     ]
    }
   ],
   "source": [
    "import torchvision.models as models\n",
    "\n",
    "model= models.resnet18(pretrained=True)\n",
    "\n",
    "num_classes = 87 \n",
    "model.fc = nn.Linear(model.fc.in_features, num_classes)  # SON KATMANI SINIF SAYISINA GÖRE DEĞİŞTİRİYORUZ\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 133,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "torch.Size([64, 3, 224, 224])\n",
      "torch.Size([64, 3, 224, 224])\n",
      "torch.Size([64, 3, 224, 224])\n",
      "torch.Size([64, 3, 224, 224])\n",
      "torch.Size([64, 3, 224, 224])\n",
      "torch.Size([64, 3, 224, 224])\n",
      "torch.Size([64, 3, 224, 224])\n",
      "torch.Size([64, 3, 224, 224])\n",
      "torch.Size([64, 3, 224, 224])\n",
      "torch.Size([64, 3, 224, 224])\n",
      "torch.Size([64, 3, 224, 224])\n",
      "torch.Size([64, 3, 224, 224])\n",
      "torch.Size([64, 3, 224, 224])\n",
      "torch.Size([64, 3, 224, 224])\n",
      "torch.Size([64, 3, 224, 224])\n",
      "torch.Size([64, 3, 224, 224])\n",
      "torch.Size([64, 3, 224, 224])\n",
      "torch.Size([64, 3, 224, 224])\n",
      "torch.Size([64, 3, 224, 224])\n",
      "torch.Size([64, 3, 224, 224])\n",
      "torch.Size([64, 3, 224, 224])\n",
      "torch.Size([64, 3, 224, 224])\n",
      "torch.Size([64, 3, 224, 224])\n",
      "torch.Size([64, 3, 224, 224])\n",
      "torch.Size([64, 3, 224, 224])\n",
      "torch.Size([64, 3, 224, 224])\n",
      "torch.Size([64, 3, 224, 224])\n",
      "torch.Size([64, 3, 224, 224])\n",
      "torch.Size([64, 3, 224, 224])\n",
      "torch.Size([64, 3, 224, 224])\n",
      "torch.Size([64, 3, 224, 224])\n",
      "torch.Size([64, 3, 224, 224])\n",
      "torch.Size([64, 3, 224, 224])\n",
      "torch.Size([64, 3, 224, 224])\n",
      "torch.Size([64, 3, 224, 224])\n",
      "torch.Size([64, 3, 224, 224])\n",
      "torch.Size([64, 3, 224, 224])\n",
      "torch.Size([64, 3, 224, 224])\n",
      "torch.Size([64, 3, 224, 224])\n",
      "torch.Size([64, 3, 224, 224])\n",
      "torch.Size([12, 3, 224, 224])\n"
     ]
    }
   ],
   "source": [
    "for x,y in train_loader:\n",
    "    print(x.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 134,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "ResNet(\n",
       "  (conv1): Conv2d(3, 64, kernel_size=(7, 7), stride=(2, 2), padding=(3, 3), bias=False)\n",
       "  (bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (relu): ReLU(inplace=True)\n",
       "  (maxpool): MaxPool2d(kernel_size=3, stride=2, padding=1, dilation=1, ceil_mode=False)\n",
       "  (layer1): Sequential(\n",
       "    (0): BasicBlock(\n",
       "      (conv1): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "      (bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      (relu): ReLU(inplace=True)\n",
       "      (conv2): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "      (bn2): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "    )\n",
       "    (1): BasicBlock(\n",
       "      (conv1): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "      (bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      (relu): ReLU(inplace=True)\n",
       "      (conv2): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "      (bn2): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "    )\n",
       "  )\n",
       "  (layer2): Sequential(\n",
       "    (0): BasicBlock(\n",
       "      (conv1): Conv2d(64, 128, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)\n",
       "      (bn1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      (relu): ReLU(inplace=True)\n",
       "      (conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "      (bn2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      (downsample): Sequential(\n",
       "        (0): Conv2d(64, 128, kernel_size=(1, 1), stride=(2, 2), bias=False)\n",
       "        (1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      )\n",
       "    )\n",
       "    (1): BasicBlock(\n",
       "      (conv1): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "      (bn1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      (relu): ReLU(inplace=True)\n",
       "      (conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "      (bn2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "    )\n",
       "  )\n",
       "  (layer3): Sequential(\n",
       "    (0): BasicBlock(\n",
       "      (conv1): Conv2d(128, 256, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)\n",
       "      (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      (relu): ReLU(inplace=True)\n",
       "      (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "      (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      (downsample): Sequential(\n",
       "        (0): Conv2d(128, 256, kernel_size=(1, 1), stride=(2, 2), bias=False)\n",
       "        (1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      )\n",
       "    )\n",
       "    (1): BasicBlock(\n",
       "      (conv1): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "      (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      (relu): ReLU(inplace=True)\n",
       "      (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "      (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "    )\n",
       "  )\n",
       "  (layer4): Sequential(\n",
       "    (0): BasicBlock(\n",
       "      (conv1): Conv2d(256, 512, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)\n",
       "      (bn1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      (relu): ReLU(inplace=True)\n",
       "      (conv2): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "      (bn2): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      (downsample): Sequential(\n",
       "        (0): Conv2d(256, 512, kernel_size=(1, 1), stride=(2, 2), bias=False)\n",
       "        (1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      )\n",
       "    )\n",
       "    (1): BasicBlock(\n",
       "      (conv1): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "      (bn1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      (relu): ReLU(inplace=True)\n",
       "      (conv2): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "      (bn2): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "    )\n",
       "  )\n",
       "  (avgpool): AdaptiveAvgPool2d(output_size=(1, 1))\n",
       "  (fc): Linear(in_features=512, out_features=87, bias=True)\n",
       ")"
      ]
     },
     "execution_count": 134,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 135,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "ResNet(\n",
       "  (conv1): Conv2d(3, 64, kernel_size=(7, 7), stride=(2, 2), padding=(3, 3), bias=False)\n",
       "  (bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (relu): ReLU(inplace=True)\n",
       "  (maxpool): MaxPool2d(kernel_size=3, stride=2, padding=1, dilation=1, ceil_mode=False)\n",
       "  (layer1): Sequential(\n",
       "    (0): BasicBlock(\n",
       "      (conv1): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "      (bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      (relu): ReLU(inplace=True)\n",
       "      (conv2): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "      (bn2): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "    )\n",
       "    (1): BasicBlock(\n",
       "      (conv1): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "      (bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      (relu): ReLU(inplace=True)\n",
       "      (conv2): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "      (bn2): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "    )\n",
       "  )\n",
       "  (layer2): Sequential(\n",
       "    (0): BasicBlock(\n",
       "      (conv1): Conv2d(64, 128, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)\n",
       "      (bn1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      (relu): ReLU(inplace=True)\n",
       "      (conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "      (bn2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      (downsample): Sequential(\n",
       "        (0): Conv2d(64, 128, kernel_size=(1, 1), stride=(2, 2), bias=False)\n",
       "        (1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      )\n",
       "    )\n",
       "    (1): BasicBlock(\n",
       "      (conv1): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "      (bn1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      (relu): ReLU(inplace=True)\n",
       "      (conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "      (bn2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "    )\n",
       "  )\n",
       "  (layer3): Sequential(\n",
       "    (0): BasicBlock(\n",
       "      (conv1): Conv2d(128, 256, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)\n",
       "      (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      (relu): ReLU(inplace=True)\n",
       "      (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "      (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      (downsample): Sequential(\n",
       "        (0): Conv2d(128, 256, kernel_size=(1, 1), stride=(2, 2), bias=False)\n",
       "        (1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      )\n",
       "    )\n",
       "    (1): BasicBlock(\n",
       "      (conv1): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "      (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      (relu): ReLU(inplace=True)\n",
       "      (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "      (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "    )\n",
       "  )\n",
       "  (layer4): Sequential(\n",
       "    (0): BasicBlock(\n",
       "      (conv1): Conv2d(256, 512, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)\n",
       "      (bn1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      (relu): ReLU(inplace=True)\n",
       "      (conv2): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "      (bn2): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      (downsample): Sequential(\n",
       "        (0): Conv2d(256, 512, kernel_size=(1, 1), stride=(2, 2), bias=False)\n",
       "        (1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      )\n",
       "    )\n",
       "    (1): BasicBlock(\n",
       "      (conv1): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "      (bn1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      (relu): ReLU(inplace=True)\n",
       "      (conv2): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "      (bn2): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "    )\n",
       "  )\n",
       "  (avgpool): AdaptiveAvgPool2d(output_size=(1, 1))\n",
       "  (fc): Linear(in_features=512, out_features=87, bias=True)\n",
       ")"
      ]
     },
     "execution_count": 135,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.to(device)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 136,
   "metadata": {},
   "outputs": [],
   "source": [
    "from torchsummary import summary "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 137,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "----------------------------------------------------------------\n",
      "        Layer (type)               Output Shape         Param #\n",
      "================================================================\n",
      "            Conv2d-1         [-1, 64, 112, 112]           9,408\n",
      "       BatchNorm2d-2         [-1, 64, 112, 112]             128\n",
      "              ReLU-3         [-1, 64, 112, 112]               0\n",
      "         MaxPool2d-4           [-1, 64, 56, 56]               0\n",
      "            Conv2d-5           [-1, 64, 56, 56]          36,864\n",
      "       BatchNorm2d-6           [-1, 64, 56, 56]             128\n",
      "              ReLU-7           [-1, 64, 56, 56]               0\n",
      "            Conv2d-8           [-1, 64, 56, 56]          36,864\n",
      "       BatchNorm2d-9           [-1, 64, 56, 56]             128\n",
      "             ReLU-10           [-1, 64, 56, 56]               0\n",
      "       BasicBlock-11           [-1, 64, 56, 56]               0\n",
      "           Conv2d-12           [-1, 64, 56, 56]          36,864\n",
      "      BatchNorm2d-13           [-1, 64, 56, 56]             128\n",
      "             ReLU-14           [-1, 64, 56, 56]               0\n",
      "           Conv2d-15           [-1, 64, 56, 56]          36,864\n",
      "      BatchNorm2d-16           [-1, 64, 56, 56]             128\n",
      "             ReLU-17           [-1, 64, 56, 56]               0\n",
      "       BasicBlock-18           [-1, 64, 56, 56]               0\n",
      "           Conv2d-19          [-1, 128, 28, 28]          73,728\n",
      "      BatchNorm2d-20          [-1, 128, 28, 28]             256\n",
      "             ReLU-21          [-1, 128, 28, 28]               0\n",
      "           Conv2d-22          [-1, 128, 28, 28]         147,456\n",
      "      BatchNorm2d-23          [-1, 128, 28, 28]             256\n",
      "           Conv2d-24          [-1, 128, 28, 28]           8,192\n",
      "      BatchNorm2d-25          [-1, 128, 28, 28]             256\n",
      "             ReLU-26          [-1, 128, 28, 28]               0\n",
      "       BasicBlock-27          [-1, 128, 28, 28]               0\n",
      "           Conv2d-28          [-1, 128, 28, 28]         147,456\n",
      "      BatchNorm2d-29          [-1, 128, 28, 28]             256\n",
      "             ReLU-30          [-1, 128, 28, 28]               0\n",
      "           Conv2d-31          [-1, 128, 28, 28]         147,456\n",
      "      BatchNorm2d-32          [-1, 128, 28, 28]             256\n",
      "             ReLU-33          [-1, 128, 28, 28]               0\n",
      "       BasicBlock-34          [-1, 128, 28, 28]               0\n",
      "           Conv2d-35          [-1, 256, 14, 14]         294,912\n",
      "      BatchNorm2d-36          [-1, 256, 14, 14]             512\n",
      "             ReLU-37          [-1, 256, 14, 14]               0\n",
      "           Conv2d-38          [-1, 256, 14, 14]         589,824\n",
      "      BatchNorm2d-39          [-1, 256, 14, 14]             512\n",
      "           Conv2d-40          [-1, 256, 14, 14]          32,768\n",
      "      BatchNorm2d-41          [-1, 256, 14, 14]             512\n",
      "             ReLU-42          [-1, 256, 14, 14]               0\n",
      "       BasicBlock-43          [-1, 256, 14, 14]               0\n",
      "           Conv2d-44          [-1, 256, 14, 14]         589,824\n",
      "      BatchNorm2d-45          [-1, 256, 14, 14]             512\n",
      "             ReLU-46          [-1, 256, 14, 14]               0\n",
      "           Conv2d-47          [-1, 256, 14, 14]         589,824\n",
      "      BatchNorm2d-48          [-1, 256, 14, 14]             512\n",
      "             ReLU-49          [-1, 256, 14, 14]               0\n",
      "       BasicBlock-50          [-1, 256, 14, 14]               0\n",
      "           Conv2d-51            [-1, 512, 7, 7]       1,179,648\n",
      "      BatchNorm2d-52            [-1, 512, 7, 7]           1,024\n",
      "             ReLU-53            [-1, 512, 7, 7]               0\n",
      "           Conv2d-54            [-1, 512, 7, 7]       2,359,296\n",
      "      BatchNorm2d-55            [-1, 512, 7, 7]           1,024\n",
      "           Conv2d-56            [-1, 512, 7, 7]         131,072\n",
      "      BatchNorm2d-57            [-1, 512, 7, 7]           1,024\n",
      "             ReLU-58            [-1, 512, 7, 7]               0\n",
      "       BasicBlock-59            [-1, 512, 7, 7]               0\n",
      "           Conv2d-60            [-1, 512, 7, 7]       2,359,296\n",
      "      BatchNorm2d-61            [-1, 512, 7, 7]           1,024\n",
      "             ReLU-62            [-1, 512, 7, 7]               0\n",
      "           Conv2d-63            [-1, 512, 7, 7]       2,359,296\n",
      "      BatchNorm2d-64            [-1, 512, 7, 7]           1,024\n",
      "             ReLU-65            [-1, 512, 7, 7]               0\n",
      "       BasicBlock-66            [-1, 512, 7, 7]               0\n",
      "AdaptiveAvgPool2d-67            [-1, 512, 1, 1]               0\n",
      "           Linear-68                   [-1, 87]          44,631\n",
      "================================================================\n",
      "Total params: 11,221,143\n",
      "Trainable params: 11,221,143\n",
      "Non-trainable params: 0\n",
      "----------------------------------------------------------------\n",
      "Input size (MB): 0.57\n",
      "Forward/backward pass size (MB): 62.79\n",
      "Params size (MB): 42.81\n",
      "Estimated Total Size (MB): 106.17\n",
      "----------------------------------------------------------------\n"
     ]
    }
   ],
   "source": [
    "summary(model,input_size=(3,224,224))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 138,
   "metadata": {},
   "outputs": [],
   "source": [
    "import torch.optim as optim\n",
    "criterion = nn.CrossEntropyLoss()\n",
    "optimizer = optim.Adam(model.parameters(), lr=0.001)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 147,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch [1/1], Loss: 0.0245\n"
     ]
    }
   ],
   "source": [
    "num_epochs = 1\n",
    "\n",
    "for epoch in range(num_epochs):\n",
    "    model.train()  \n",
    "    running_loss = 0.0\n",
    "\n",
    "    for images, labels in train_loader:\n",
    "        images, labels = images.to(device), labels.to(device)\n",
    "\n",
    "        optimizer.zero_grad()\n",
    "        # Forward pass\n",
    "        outputs = model(images)\n",
    "        labels=labels.long()\n",
    "        #print(outputs.dtype)\n",
    "        #print(labels.dtype)\n",
    "        loss = criterion(outputs, labels)\n",
    "\n",
    "        # Backward pass and optimization\n",
    "        loss.backward()\n",
    "        optimizer.step()\n",
    "\n",
    "        running_loss += loss.item()\n",
    "    \n",
    "    print(f\"Epoch [{epoch+1}/{num_epochs}], Loss: {running_loss/len(train_loader):.4f}\")\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 148,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy: 74.69%\n"
     ]
    }
   ],
   "source": [
    "model.eval()   \n",
    "correct = 0\n",
    "total = 0\n",
    "\n",
    "with torch.no_grad():   \n",
    "    for images, labels in test_loader:\n",
    "        images, labels = images.to(device), labels.to(device)\n",
    "        outputs = model(images)\n",
    "        _, predicted = torch.max(outputs.data, 1)\n",
    "        total += labels.size(0)\n",
    "        correct += (predicted == labels).sum().item()\n",
    "\n",
    "print(f'Accuracy: {100 * correct / total:.2f}%')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 149,
   "metadata": {},
   "outputs": [],
   "source": [
    "torch.save(model.state_dict(), 'resnet18_finetuned_geam.pth')\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 150,
   "metadata": {},
   "outputs": [],
   "source": [
    "torch.save(model.state_dict(), 'model_checkpoint_rest_net_ge.pth')\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "optimizer = optim.Adam(model.parameters(), lr=0.001)\n",
    "model.load_state_dict(torch.load('model_checkpoint.pth'))\n"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.9.0"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}