{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "754e069a-8cc6-4b1c-b528-6632467082a9",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/opt/python/lib/python3.13/site-packages/tqdm/auto.py:21: TqdmWarning: IProgress not found. Please update jupyter and ipywidgets. See https://ipywidgets.readthedocs.io/en/stable/user_install.html\n",
      "  from .autonotebook import tqdm as notebook_tqdm\n"
     ]
    }
   ],
   "source": [
    "import os\n",
    "\n",
    "from datasets import load_dataset\n",
    "from torchvision import transforms\n",
    "\n",
    "import torch\n",
    "from torch.utils.data import Dataset, DataLoader\n",
    "import torch.nn as nn\n",
    "import torch.nn.functional as F\n",
    "import torch.optim as optim\n",
    "\n",
    "from PIL import Image"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "263afdac-5f2a-49d7-8242-9543408b2584",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training on device cuda.\n"
     ]
    }
   ],
   "source": [
    "device = (torch.device('cuda') if torch.cuda.is_available() else torch.device('cpu'))\n",
    "print(f\"Training on device {device}.\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "36353151-00f2-4c34-b6ad-38053233aeca",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "DatasetDict({\n",
      "    train: Dataset({\n",
      "        features: ['image', 'label'],\n",
      "        num_rows: 32000\n",
      "    })\n",
      "    val: Dataset({\n",
      "        features: ['image', 'label'],\n",
      "        num_rows: 4000\n",
      "    })\n",
      "    test: Dataset({\n",
      "        features: ['image', 'label'],\n",
      "        num_rows: 4000\n",
      "    })\n",
      "})\n"
     ]
    }
   ],
   "source": [
    "dataset = load_dataset(\"hadrilec/satellite-pictures-classification-ign-france\")\n",
    "print(dataset)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "297e6e52-8e85-4eb3-b4b1-4a3fd5962d87",
   "metadata": {},
   "source": [
    "# Dataset Creation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "2ec5367d-0c43-416f-a10d-715060c9be44",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(320, 40)"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "class HFDatasetAdapter(Dataset):\n",
    "    def __init__(self, hf_dataset, transform=None):\n",
    "        self.dataset = hf_dataset\n",
    "        self.transform = transform\n",
    "\n",
    "    def __len__(self):\n",
    "        return len(self.dataset)\n",
    "\n",
    "    def __getitem__(self, idx):\n",
    "        example = self.dataset[idx]\n",
    "        image, label = example[\"image\"], example[\"label\"]\n",
    "\n",
    "        if self.transform:\n",
    "            image = self.transform(image)\n",
    "\n",
    "        return image, label\n",
    "\n",
    "# Usage\n",
    "transform = transforms.ToTensor()\n",
    "\n",
    "from torchvision import transforms\n",
    "\n",
    "# Training augmentation pipeline\n",
    "train_transforms = transforms.Compose([\n",
    "    transforms.RandomHorizontalFlip(),       # flip left-right\n",
    "    transforms.RandomVerticalFlip(),         # flip up-down\n",
    "    transforms.RandomRotation(20),           # rotate ±20 degrees\n",
    "    transforms.ColorJitter(\n",
    "        brightness=0.2,                      # change brightness\n",
    "        contrast=0.2,                        # change contrast\n",
    "        saturation=0.2,                      # change saturation\n",
    "        hue=0.1                              # change hue\n",
    "    ),\n",
    "    transforms.RandomResizedCrop(\n",
    "        size=256,                            # crop and resize back to 256x256\n",
    "        scale=(0.8, 1.0),                    # crop 80%-100% of original\n",
    "        ratio=(0.9, 1.1)\n",
    "    ),\n",
    "    transforms.ToTensor(),                   # convert to tensor\n",
    "    # transforms.Normalize(\n",
    "    #     mean=[0.2918, 0.3070, 0.2985],           # same as during training\n",
    "    #     std=[0.1065, 0.0985, 0.0877]\n",
    "    #     # mean=[0.485, 0.456, 0.406],         # standard ImageNet stats (optional)\n",
    "    #     # std=[0.229, 0.224, 0.225]\n",
    "    # ),\n",
    "])\n",
    "\n",
    "# Validation pipeline (no augmentation, just normalization)\n",
    "# val_transforms = transforms.Compose([\n",
    "#     transforms.Resize((256, 256)),\n",
    "#     transforms.ToTensor(),\n",
    "#     # transforms.Normalize(\n",
    "#     #     mean=[0.485, 0.456, 0.406],\n",
    "#     #     std=[0.229, 0.224, 0.225]\n",
    "#     # ),\n",
    "# ])\n",
    "\n",
    "\n",
    "train_dataset = HFDatasetAdapter(dataset['train'], transform=train_transforms)\n",
    "train_loader = DataLoader(train_dataset, batch_size=100, \n",
    "                         num_workers=10, pin_memory=True,\n",
    "                         shuffle=True)\n",
    "\n",
    "val_dataset = HFDatasetAdapter(dataset['val'], transform=transform)\n",
    "val_loader = DataLoader(val_dataset, batch_size=100, \n",
    "                        num_workers=10, pin_memory=True,\n",
    "                        shuffle=False)\n",
    "\n",
    "len(train_loader), len(val_loader)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "f909a55c-89ab-4d1c-894f-8a641350e219",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/jpeg": "/9j/4AAQSkZJRgABAQAAAQABAAD/2wBDAAgGBgcGBQgHBwcJCQgKDBQNDAsLDBkSEw8UHRofHh0aHBwgJC4nICIsIxwcKDcpLDAxNDQ0Hyc5PTgyPC4zNDL/2wBDAQkJCQwLDBgNDRgyIRwhMjIyMjIyMjIyMjIyMjIyMjIyMjIyMjIyMjIyMjIyMjIyMjIyMjIyMjIyMjIyMjIyMjL/wAARCAEAAQADASIAAhEBAxEB/8QAHwAAAQUBAQEBAQEAAAAAAAAAAAECAwQFBgcICQoL/8QAtRAAAgEDAwIEAwUFBAQAAAF9AQIDAAQRBRIhMUEGE1FhByJxFDKBkaEII0KxwRVS0fAkM2JyggkKFhcYGRolJicoKSo0NTY3ODk6Q0RFRkdISUpTVFVWV1hZWmNkZWZnaGlqc3R1dnd4eXqDhIWGh4iJipKTlJWWl5iZmqKjpKWmp6ipqrKztLW2t7i5usLDxMXGx8jJytLT1NXW19jZ2uHi4+Tl5ufo6erx8vP09fb3+Pn6/8QAHwEAAwEBAQEBAQEBAQAAAAAAAAECAwQFBgcICQoL/8QAtREAAgECBAQDBAcFBAQAAQJ3AAECAxEEBSExBhJBUQdhcRMiMoEIFEKRobHBCSMzUvAVYnLRChYkNOEl8RcYGRomJygpKjU2Nzg5OkNERUZHSElKU1RVVldYWVpjZGVmZ2hpanN0dXZ3eHl6goOEhYaHiImKkpOUlZaXmJmaoqOkpaanqKmqsrO0tba3uLm6wsPExcbHyMnK0tPU1dbX2Nna4uPk5ebn6Onq8vP09fb3+Pn6/9oADAMBAAIRAxEAPwDiLbQ9HiDC308zs3XzBuxj0zVqHTVtpN0NhbWj4x5m0A49KJNQndQm8KM5+UYNVnmaT77s31NOwXNBo2AzcXq7O6x9fwqJmsV5BnlP9124P1qgWyelJk01EVy+L2NB+7s4Vx3Iz/OmNq16ykNcyf8AfVUs+9NOKfKFxxbtk4o3du3pTCfSkp2JuO70nWkop2AOlKDSUUWELn0o3GkpRQMXNFJS4osMOaAKWkJIOMHNAC45pQpY4UZNTC1YJvkdUHp1P5VOI5BGGiURR4/1kg2lqlsZGlvbpj7TOy56og5q5GxK7bNFih6b5fvGqbTwxkCJFkf+KR89fYVHJNJI25iSfyqbNjuX5Tar/r5JJ5B3U4XFNF3bx8w2yjPXf81ZxkJGOlNGTT5e4XNFtUnRAsbiMDr5YC5+tQyahcyrtaZ2XOcE96p96Xp7U7ILkzzSyrtd3YehYkUzBHsKYM4OTSHBosK9xxCnvS7kA6ZqPAHejg0+UCpqqpLplyGXOELfiORXE13lxGksEkbZ2spBxXB1MkCPRSw7Um/mm5pO1WIdkmkpKKYgNBoooATFFGM0c0gCijFLigBKKXjPNLtycKNx9hRcLDRS1KltK3Vdg9W4FKYlX70qnHZeaLgQinKCTwCT9KmQLuBSB2PYkcVOElYfPPDEPY/4UrjIfIKcyOi/jmpVJA/dQYH95qb51koISJ3PqxqvJcSSMcsQvYelLcCYNEjlpR5sh6+gpkknnNuZjgdB2FQGk696AJC2OlNJZjQq/U1aitiULOCPQetFwsV1Vm+6Caf5MucH5asxqqHDHA9qS5kjtxvVS5PY0r3GQeQe7flSeQM8k4qu1/KzHCAH0xTZZ7nGSMD1pibLn2cAZA6UhjHOCDiq8cu6I7nI/rUIk2ZZX69qaRPMTGaLJHeq6yOZMcbc9KJEBkBGee+KZIuxwAeatCbJ5HIJQjjHbtXDSp5Uzx5ztYrn1xXb20i+a3mN8rDkVx2oRmLULhSP+WhI+h5H6VNRFQZ3Aoqb7LL6ADudwpDBjH72P3wT/hSuMiFFS7Ij0m/8dpypAfvGQfhRcViCirYiU/MltKy+uetKDKvCRwoPR8E0XCxVVHPKoT+FSLbTPztCj/aOKlZ3zl7jYR2jHFMZ4SRvaSU+ucUrjD7OirmWdUPYDnNJshUcb5fXAxigTRJ/qoR15380jXMh4GEH+yMUwJFOB+6tgpP8UnNIzyL951UH+4OahaQlcNISPTNM+UdjRZgTM6cEs7+xpvnBQdsaqexqI+wozRYTZI1xKwwZG+gNR/gM0UcdapJBcKOaKjeQqMDk0WEPZlXqaRbhNwUfrVdo5m6qaVLeVlLbDwcYoaQXND7VFEM8Z9qX7e8gyinNZbKwOCCCKmtZHWQBMc9c1LjoFzU25VSeWIzmq5BLu8jZA6AVciUunAy/aoJ4weEPz96ziyUyqQWmCkDpnPpViGRkhkDENk5GBnJqjNckNhRg/wAXvUbXb7Nq8L6Vpa42TTNGdyqQOegqsTgYqLJpQecGrUbCJ0dyOCeO1TzBTZhyPmzVVJCp4HHrVvzFZBF2Izn1pgURjPIrm9VVl1GQkY3YI9xiurkiQKpT05PvXM65/wAfy/8AXMfzNTPYqG53O+0AJ2SE9gTxTftEHObZfzNVsilyMVmUWBeELjyoyP8Ado+3XGPv8H2qvnijPFMB5mkYnLnn3ppPcnmm5FJkUBccTSLS7sdqaC2eBQF7hzmnUHOO1Nz7imgF+popufnxg5AzTnBWISKpIPTimIaSMZ6fWmecnrkmoXZ356DsKVEXOcE4PamhNkjy7PvKaZ57f3a0kCzJhlB/CklSHaMoBtHalzaiuZvnsevSpowTGehDHp6VWbGSB0zxUyg7AQDxVNBcsoArKQxz0NXQwJAGOBzVNPnwUxkDNTxkAeYWwe+azYmOmtVkAI696pC0kR8gcE8Yq+SXU7Tz3pqgyPuVgNpwQaVxJj0uY0jwWwy8EetPYeYyuAApXJxUf2dJjjG18kE9j9DUc87W8ZVQCNuMZ6VCWoXM++hEc+Bn5uaqhT6Vrpbm8ieSYfvl6ENxUC2LbhHzknk1snZFFEjimc+lalzaBJmZsLEOhxWew5OOapMQiHBHPHepGfcxxnA6GowParNsf3gBHy9CMU2BJE6tA24Z4rmvEO0vblRgYb+ldh9nRWk8sKQOlc14l2G3jKgcSY6dODWb1RUdzZxml2tirCKhX5l6jimzukS565NIdyLaelG361PHl0XJXaelNjkAVsAMFOM0Bcrh15ODgd6aZVUZIP5VPhGkGBx3FS3KRm0Pynj+7T0E2QxujAEKSR1q2YUeRdo4PJrPs2Kg7cEkZ5NX1kQy7s4B7evHSploybjE8t7kpjgKc+lJNbRRvG+Sd3YVMIAY2wgB/hXPPTmq6lnJyxJUcDHSkmNXIrglG+RTxwTVm0JlUoflx61Sn3Bjh87uSBVgO1uwJIYuBVvYGOktwCzSEGM9MdqZZ5lk8pVG0dTVh2Bt2ABYGq9s3lsJIzluQy+1TcViQArJleADj61JJhkPOAe9KyozFgchRnjtmq0jDyChPI6D0oW4J3KMoxKwBzg07OBnJ9hUWc9akjCscE4PathlmJGyARjPXFWZPmUggFc8VV3bAQGO4Cp7fLkvk89qyYixIwiiDAdewqrZ3QjkbzR8rH8atuu8gFgAvQVWuY/Mw6gDbweMUgSL5lHm+bEQyEcKOlYkjMWO4YJOMVdt1kjmVCPlPOVPekbT5HkZiMDPHvRGyCKC2mfzRsUheAa0pJI4mQyHAP61BHD5e0A4x1p91EtwEHK7B60PUrlI5GDYilIIAzisl0yxIHGeK1RaKSGZiW6k561KIowANoqk7AomVBZmSMuTjFEMak/fIAPJrX2ADbxj2qP7PFjGwYzmnzlcpWLlofLgU/X2rn9bgzp0jMDlCGB9Ocf1rrdoHQAfSs/WbeObRrtSMYjL8eq8j+VTfQFFIZ5pVoeOCcGpriJUBQclhwKqRyHgMTweK0FPnhC5xjpQyCKJNkarjPHX0NVbgeSxVGB3ck1pSoyOpi4XofQ1RuVS4bdGMEdQBSTENiUNt2NnHLVZBLFh0GMAetQpGixMADwM5pgnjZA33SBwPU07DKzHdLgfKQcHFXXQssZ7qgGfUg1SVg7MzdauWZZyFZTjOR60T2EXH3KzOD8pBK+54qAxPJIGRsK3Wi6BW7GT+6UcDHeleQC5wwYRjAQYwBUILEAkUSGJlUsM4PvTfs+6MBsmT1zUsNoJHDM25s5/CrNzbICRuKsRkYp3ArNIYUADZUEAio9mJPMQbiTwB2pscDTO0YcDuA3c1ahyZTGUKsF4z0psGLHKv2nbkZK4YepqC4UQMSFJHc1A6mO7Lv8AeDc4q7dSB1ibYPKVwWGepoCxmrFlvunJ7U+OGNolLthy1WHX/S8F9yE/eHHB7U82WJnAwoHIBPWr5gK0jDdgDJHf1q/Yx742/hI5pUsgsW7I3H1quJyd0a8Dodvep3GtdizclAUX1OSR3qzFtlYBl4A6VQj3yhVZchTwavQwlSGJOO/NQx2HvbrDNnruH5UhOcYp8zrIw2ggD1plCRcUJRRS0DEopaKYCUYpaKAExUF5AbmyngUgGSNkBPbIxVnFNoAwBv3DAzz0rRhzt3nnHGKpW4IkBz+FXYgDIc5HtmrkZstFlKqM5wM1FEotY5ZiRtPao5FeFQ6rvjzhh3qtPe77ZkHCscAVNrkskursG2RowAW4YVmZyB7dKBuIwMnFG0jg1rFKwDl45rStSXAJOMcACs1VJI4q5bEiQYPfpUz2GjWKAL+8G5m7U+IJO8kTjKr0qvcXCOVVW+Ydfb8anggYyIVbA7+9Y2AzoGMV7JtUmPO3d6VLezSMrMqfulO3eKbcSu91PHFja55bvT222ts8ZXcXXHXpVJEvcq2xCXBEZ3Kwxmp55dh2RklgDye1UQku0bcjHTFXrGB5bjnlsZwTTaK5SCDmQq4LBmBbir09tEUEeM5bIpixFrhifl5q2SOCB0FJsrluil5PmMRgKB1qeFCh3k54xzUmM/U0lLcfIh0jB0CkVClvGjl1HNPZgi7mOAKge9iUZBzmmkxqy2LPtjmlA9TUUEqS55INXBAHj3I3PoalvUaaIPxpM07YeTjI9aOKdx3ExRRRQIKKKKooKKKKADmkxS0UiTChUSPjdj0q0qSh8qBj69apRqT8w+6OtaEUTS+XtOOO571cjMtRkFR8pwP51k6jbpFIrKwwf4fQ1syfJCxCnPUAd6wLmZ5piWB4OAD2qae5I6Ahvlzgmpm+QhWQcjr6VXjyZVGKtyh55c7guRitG7APWKJIlzyW6YqPyJEOV4BNXorBoIQHOVJB3CrQtwI/lJJz0rNzC5mxwsrcg81oSXLxLtRcnHWmzAQoHY8EflWabtucdM04q+pWjHmCSM71PXn6U+PLk72G4+tVmnZge+aiy687q05R3LzTxRtsJ5HXFWI1RpUnyfl7dqxSx359avWcjq+CSQaiUbCuWJro+YNgwc9K1rCaP5RJGMOfmzWTdWoedSgzxkmrST+RbKjAEsQBUNDTNY2EE8kjo52jjgdKxrmT7MG3D5h0BrUtrmOzEjSvtQj9a5e+uRcXTMrFl9+9EEO4jzS3GR1HoKhIxg1NZqjyfMTnHSnzW7RkMuGUHpWt0K5NYOoY5bn1NaEd5sJKYKHjNYW7fIC42/StKMQtCqRt8+elZyhqKxp27HZkkbTSvCPNB/hIqn9p+xuiyKSjdas3MoVQwNQ00O5I1md4VGzxz7VHNA8LEMMe9TRSlgrgZBAq5Mi3cO3+IcipvqNMyO1A6UrAqxB6g03tWhYtFIKWmMTvS0UUAYSsWZVAwM9K0Itw3bTwPSswc4Jq/bsC52g4btVSMTQ3+TC25CQBxjrXPTSRvIWVSD3B9a3YzkPzkIO9Y/kSTyOxGCSTShoSVhIVJI+9VmBjvR5MhD696q7WD4IIPpitO2kMypG0QIj6fWqkD0LT3ZLhI8FADxU1vdGQsTEVUdW9DVFxFAjHePO9BVmyk81FEQA2glyeprJx0FYt3GLuwYxqFOCOa58ITkcjHOTXQR3AjRBEoCN1rN1FAs27Iw/anSuhopqyooHf1xTJHBPFSF128ZyKicZO410jGFSSCKuQhwRhhVYcg80+NyrAgZxUSGXGDllidjk9MUpilYbSBkdDmmKzI6zNgVft7mGa5BUFh3rMLlK/nlltAsi7TwPrVCFRkEgcVa1CVTKVXnBxnNV0KrGfWrWiAnT/AI+VwuAR1rTlAjgQ8AnqTUNj5Zt8ygbgeKddF2CnAVWOKzb1FczhF5h47nrWpHZtHbb1A3qMjA5NIsK+WHiU7h14pl5fIgbB2zBMdOtNyuFylPfPOu1k2leueuaka5M0oUn5eMflWaWcjJ7nrT0kIPNU46DudJaNhB83Tgir/wBoS35Y4z0rAtLhlj2ZwW6HNSnEhABOV5wT1rFx1JvqXZlzmTIwaiz+NNmuiIx8oA6Gmi4/dDAyvdqdmaKWg/3opEZXUkHtxVS4nmiTeACp4+hppBzFssAMngVDNdxRMFLZzVRLhpbdlLfMT09aq3MZVzkEGrURORBZFpLWCR2ySikn1OK04mCLvXAz0zWNpLmbT4SeMLt/Lj+laUabyBnpQxMt5YWxfo45OOlRQy5G5jtz2ouJl8po+Q2OfeobaJXyrsfalYRoQwB5C/DsoytVZLg6deFgobcMn0rQ05CI3BOT2qhfhxMwJAHpSTuLczpWeSbzmx83JIqa2uHiJAJ+YEfhUlum59uOO9WDZqJiW4X2qm1sMms3WZ0UkjaOarapG63Rc8qRwe1W4lEbBsYXsaoajc+bhA+SG5xSjuIqeZj6U0sWIHOKm8kCMktmljzFn5A4I7VrcZJFbIyk7vyqSO3KSbHfAPRhVWOZowdpwDVi1fEg3HOaiQi1cRR+WqO49mrNMkltujR+vcVpXMZkUnsBxWRhmYZoihkskDpGHY5J60sADH5hn2pXaR8K3ReKdBKkU+7qKpiLBgaIiRG+UdjU5kMli/mj5gflJqvc3kbIAucmkUNtVnYlfSs2gLKXhRAgJ6dabcxmaJmYA8ZB9KZPBmRGiBw3WpJZWgt1UqG3DHIpWBFSAqwEUi5Ham3Fm8KeYfu7sU63QkgjrnP0rSnkZ7fY655+Y/hxVtjZlRAYXa2JCcAZ4FWWkMM2A2WHWop7VYJVAYsCART5IHlQSjsOlG4JDijXETuXXPXGafaSSFDGAB254qGJjCxU9Mc0+2d5Zd4XLZ/IUmhj0aVXZWXAU0XuDZgkHOaknmiluQASMdR71VuppZf4cIO1KO4EdtErAsSRjpUlygeTnOMYxUURcDG75Sav3LILddnccmqvZgcxobNLY7AMbGIz69/61tW6ZPU8Vg+HZQsMqDqGyfxH/wBat+2LLkYzkUugPcW/ZdyjHzKO1RQs+flOKL0hmUjg46Glt5fLXLqSKFsJbGtazbvoo5x3pk8SXb7lPzL2qK2IfJUEZ9KsQqyb+cgHHuah6C6laCPBZmxgHrUjbzcLnBXaDmkh2pMxYDDE/J7VA7lJG27cE8AnkCklcW7Jr+7AcRgYUAZGKx5CpkJUYFWJHEkgLctirBgiax3hcEH861WiKKwB8hU29ea0AsUFpuK7mbsKpwuI4mYnnsDVzTMO7F+3PPak2IoRwqxZj8q9qns7ffJkEcc027kAuGQEFc54pI5GXIiyCRyaALDF5sxx84B/Go2hSG2Khl3Zyc9RUlg6JvyfuAn61UumkmmadlwrUK4Eq7pcIwxk/eHeqdxEYZioYYHpVqW5HlxpHwy96pMS5y3JNWkMVULtgZJ7VZiJOUYEn0qujMjAqeatQIf9YzYPWkxGgJdkYYfQLTJlJ2hlJyOPan4DR7iMFhnmq013vhIGdw4BrNAgh8tLgbSWTofrVm6cRrGoPy9cHtWVDKYZQQCwHOKdPK9zLkA47VdhsnuZUmmyoyAMCrcIC2+/sOoqhCdpAYc1ZkneUeWg+U+lDDYFeL94SR0PJqpb3BgLFfvEYBPapPszEhepP6VJHY4HzdaV0MjiJeYM43Z64rQuEUwMGxnFQw2pD7jwBVidDKD9MVPMgM6Fkif94pKnsKS8JkkAVWVcVdS2UAZOSKl8tSdx60+YDhfDzqJZkJ5IBH4Z/wAa6y1KhSTy3tXH+H136iwJ/wCWZ/mK62LCSphCwHWjoEtyG73edlxx2qW2lC8OMr6YqKZXklZ8EDOMHtU1vCwcbqd9ARetxGjbowck8gmpcMsnz+vao8eXlsYJ6U6SVI4DM3Lr2NQ1cloqXMggucKAMckmqEsvmS71/HFNup2uZjI3foBTY8bvmB2961S0AnEEpUOFPJ61ZEwt4jC/zAjOB2pqXKxweVHJtzzk1Vl6hyxJPU0tw1HQyojsWXI9KfDI7SkrlQaZFCztnb1FW7WOQMVYDHWh2SAm+zRvHyoJqCGVFlfcoG3pVpkcHA6VCtkNx385qUwKLtGochiXY8YoaZ5o1j9OMYxVuWwG9TEOO4q15ESqo2jim5JAZU0JQlOS4P3QKelk5j3Nx9K1mVWIZutLwOKnnYFFrOIwrtDbgeSe9SxWsYCk5J7jtVjAwfrSKvJNJyYCMgdNjdOnFNWGNeijFTHGOKQDB9qm4EXkodxC4zTPsyj7pwasduB3oBwTVcwtSsLVWkyT8vpVgRIo4HNHelHWlzMd2CgA9qccdaacZGKUtzilcaF5NI2QMdaXOTjFBB3ZFIYdM54oyOgpChznNGDtyOtAHnWgqG1q3B7lv/QTXoEMAVcnCkdPWvN9Oz/adrt6+cmMfUV6fBwcS8emaq45AbcJCCcHJyaiDHn5ePpUskpWMID1/SmBuNuOaLkjTlivtyKJokmb94M59KUg4NIcjBxzRcRAbOIxlQvHrUIsMtg8IPzq9kZyAcCjOVo5mO5T/s5Gc9k9qmFlEUKtzn17VNnjFGeOetHMwuIsITAx0FOGA2MYzRk9c0ZGOetK4XDgg8ilXaOppjEE9MUmOaQDmcBuO9IRnnNJtyc07+H6UCFcqTxQSMBcU3gDNB9qAFDAcY4peBTQPQUYOOaAHZNNPtTsY6UBGJ6cUAN5xSgd6f5bDjFLsAAOaLgRYI5ByKcpyORzUmxeTnPFAUY4FAERBzTtuR1p4X+8KXp6YoBDVXI78U7aTj8qXPB60EqPXrSGNZDSrH8oyQDRvUjpnFKH9hzQM8nikeGVJYztdGDKfQjpXqbSsYweNw9q8pr0qwnkubC3lkxueJWOPXFUwmTHnqe1KM8Cl2EHPWl44x170jMTGAeeaUAjkGlwcUoU7aBoQkjg96aflXPUVIPek2YG4njPSmMZ+NKO3enBMH7ufrT+BjGBSAZtz7U7YuMbjSggDrk5pdw7Ligdhu0Yo2D3pd2WoDN68UBYTYvoaXYCvANLkkGkO7BA+uaAF2KVI5puzA6GnHcBnPam8MoOaBCrgfw/nTjyOMCm4GfXig8rxQAqn1/SnhgATTDjbgdxSZD8Y+lIBzNj3NJvweAKYcg4NJ3/AAp2AfvbPak3k4JzxSqoBAz1pxA2nB607AMySOtAOOMUpKkqMYxSEgn09KLAGcGgnc3y+lH8QpV9aB3EOAo7ZpQSAMdaTHOfSkzznv0NAjymvRdAk+0aJavt27U2deu3j+ledV33hWaN9DjUMSY3ZWHuTn+RFDNJbGwQARkcUoGM4X8aduJ4GMelIQXG3OCOaRmKM44YCkwOuckUnsByOaQHnIoAXcoICjk+tLvboMA5phwwYdD1FBBKq+fmoAcxJyrHOKZzjOelPwBgk9etIAXJAHGeKaGhNuQDnGetOO0gnrSsMAfWlH3utDHYTgHNIx54xTjjG0c+pqMcDd3zxRYB6nORnHTNLtwSOaXeMfdwD3ppkIHsBRYAI4A60hUZwDwKM8cH3pp+9nNAhx4PHGe9LwOh6U3PyHnI6UHbxQIcxwy/TpSRvgEjrmiTO7j0oU8dOMU0UIXJ4xzRhs4zninKrZOAMmk2sJOeMDmkIaoJYc0o2kHORig4QAnuaViCMYxzjNVYBMj0yaN3IG3pQTtY07cOoHNFgECkkkjilIwP5UgkyhHfNOL5bpjsKVhCY29aaVOQfbJpWbfnjkUZQYxnODRYDyiuy8Gyp9jnj3DcJNxHsQMfyNcbXUeDhl7znsn9aGaT2OvxtPHXrTwQZRj+IcZqLduOP1pQ/wC9UnjAxUmSAqQT7+lCADlTwO1NkJ3g9s0EZLYyM9KLDuPIXcR3PNMB/d/KfenEAspbgkYpqphSvcZqrALuyoLDpSj/AFmAxAPNIrZjxxSleh4yRSAX5STycU1j8o4pwADcnjFKGXBGKdhiKRlRj6047SG44zxScA56AcZpM5I447YosAoBxjOKBHwcnp2FKxLfwU3LBeeppCHBABzTVUHk/lSlyQM4AppBDZz9KAJAkYByT16UFUB4bIz1qMn1PWlACANjr2oAVwCSc9sChHATBHPvQfuKccg004D8j73SgoeJO+3kU3L4wecmkAJDfMOnFVP7V07/AKCFv/39FAFwqSpHvxRnK4PHNZEniTSo5GX7UTg4JVCQfocVVk8YWAchYZ3A43YGD79adwszoSRtK9aUOvAIzXJS+MsSHybPKdi7YP6VUm8XXzyZijijT+6Ru/WgfKztmx0Cde5NKP8Aa49K8+n8S6pOVPn+Xt7RjGarT6tf3KhZruRgDkc4/lQHKekSOkaBpZFQdMscVBJfWUETSyXUW0f7QP8AKvNHmlkGJJXYDnDMTTKA5T//2Q==",
      "image/png": "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",
      "text/plain": [
       "<PIL.Image.Image image mode=RGB size=256x256>"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "img, label = train_dataset[100]\n",
    "pil_img = transforms.ToPILImage()(img)\n",
    "pil_img\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "db4e313d-ca78-48fa-b105-48e4059c9577",
   "metadata": {},
   "source": [
    "# Model definition"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "e673efc8-7749-41bc-9f60-885c83490509",
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "class Net(nn.Module):\n",
    "    def __init__(self):\n",
    "        super().__init__()\n",
    "        \n",
    "        # Feature extractor\n",
    "        self.conv1 = nn.Conv2d(3, 16, kernel_size=3, padding=1)\n",
    "        self.bn1 = nn.BatchNorm2d(16)\n",
    "        \n",
    "        self.conv2 = nn.Conv2d(16, 32, kernel_size=3, padding=1)\n",
    "        self.bn2 = nn.BatchNorm2d(32)\n",
    "        \n",
    "        self.conv3 = nn.Conv2d(32, 64, kernel_size=3, padding=1)\n",
    "        self.bn3 = nn.BatchNorm2d(64)\n",
    "        \n",
    "        # After 3x maxpool (stride=2), 256 -> 128 -> 64 -> 32\n",
    "        self.fc1 = nn.Linear(64 * 32 * 32, 256)\n",
    "        self.fc2 = nn.Linear(256, 64)\n",
    "        self.fc3 = nn.Linear(64, 4)  # 4 classes\n",
    "\n",
    "        self.dropout = nn.Dropout(0.5)\n",
    "\n",
    "    def forward(self, x):\n",
    "        # Conv layers\n",
    "        out = F.relu(self.bn1(self.conv1(x)))\n",
    "        out = F.max_pool2d(out, 2)  # 256 -> 128\n",
    "        \n",
    "        out = F.relu(self.bn2(self.conv2(out)))\n",
    "        out = F.max_pool2d(out, 2)  # 128 -> 64\n",
    "        \n",
    "        out = F.relu(self.bn3(self.conv3(out)))\n",
    "        out = F.max_pool2d(out, 2)  # 64 -> 32\n",
    "        \n",
    "        # Flatten\n",
    "        out = out.view(out.size(0), -1)\n",
    "        \n",
    "        # Fully connected layers\n",
    "        out = F.relu(self.fc1(out))\n",
    "        out = self.dropout(out)\n",
    "        \n",
    "        out = F.relu(self.fc2(out))\n",
    "        out = self.fc3(out)  # logits, apply CrossEntropyLoss\n",
    "        return out\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e2361c26-a0a2-44d3-8f25-c7155f50d558",
   "metadata": {},
   "source": [
    "# Training Loop"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "665c66e4-903a-4e4f-b760-547561c8dba5",
   "metadata": {},
   "outputs": [],
   "source": [
    "from tqdm import tqdm\n",
    "import datetime\n",
    "\n",
    "def training_loop(n_epochs, optimizer, model, loss_fn, train_loader, val_loader=None):\n",
    "\n",
    "    history = {\n",
    "        \"train_loss\": [],\n",
    "        \"validation_loss\": [],\n",
    "        \"epoch\": []\n",
    "    }\n",
    "    \n",
    "    for epoch in range(1, n_epochs + 1):\n",
    "        model.train()\n",
    "        loss_train = 0.0\n",
    "       \n",
    "        for imgs, labels in tqdm(train_loader, desc=f\"Epoch {epoch} batches\"):\n",
    "            imgs = imgs.to(device=device)\n",
    "            labels = labels.to(device=device)\n",
    "\n",
    "            outputs = model(imgs)\n",
    "            loss = loss_fn(outputs, labels)\n",
    "            optimizer.zero_grad()\n",
    "            loss.backward()\n",
    "            optimizer.step()\n",
    "            \n",
    "            loss_train += loss.item()\n",
    "\n",
    "        avg_train_loss = loss_train / len(train_loader)\n",
    "        history[\"train_loss\"].append(avg_train_loss)\n",
    "        history[\"epoch\"].append(epoch)\n",
    "\n",
    "        if epoch == 1 or epoch % 2 == 0:\n",
    "            if val_loader is not None:\n",
    "                \n",
    "                model.eval()\n",
    "                loss_val = 0.0\n",
    "                with torch.no_grad():\n",
    "                    for val_imgs, val_labels in val_loader:\n",
    "                        val_imgs = val_imgs.to(device=device)\n",
    "                        val_labels = val_labels.to(device=device)\n",
    "    \n",
    "                        val_outputs = model(val_imgs)\n",
    "                        val_loss = loss_fn(val_outputs, val_labels)\n",
    "                        loss_val += val_loss.item()\n",
    "                        \n",
    "                avg_val_loss = loss_val / len(val_loader)\n",
    "                history[\"validation_loss\"].append(avg_val_loss)\n",
    "            \n",
    "                print('{} Epoch {}, Training loss {}, Validation loss {}'.format(\n",
    "                    datetime.datetime.now(), epoch,\n",
    "                    avg_train_loss, avg_val_loss), flush=True)\n",
    "            else:\n",
    "                history[\"validation_loss\"].append(None)\n",
    "                print('{} Epoch {}, Training loss {}'.format(\n",
    "                    datetime.datetime.now(), epoch,\n",
    "                    avg_train_loss), flush=True)\n",
    "        else:\n",
    "            history[\"validation_loss\"].append(None)\n",
    "                \n",
    "    return history, model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "68fcd54c-4691-4905-9ad5-725d335a03d3",
   "metadata": {},
   "outputs": [],
   "source": [
    "# build model and move it to the GPU\n",
    "model = Net().to(device=device)\n",
    "\n",
    "img, label = train_dataset[0]\n",
    "img = img.to(device)\n",
    "out = model(img.unsqueeze(0))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "f157c1e9-0c1b-420b-99f5-a8f546d9bfbf",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "torch.Size([1, 4])\n"
     ]
    }
   ],
   "source": [
    "# four classes are predicted, output should be aligned\n",
    "print(model(torch.randn(1, 3, 256, 256).to(device)).shape)  # should be [1, 4]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "aa4507e6-57f3-469d-a899-be5db754ed51",
   "metadata": {},
   "outputs": [],
   "source": [
    "# stochastic gradient dissent is used\n",
    "# learning rate is 1e-2\n",
    "# we want to get the probability that each picture corresponds to each category\n",
    "# consequentlym we choose cross entropy as the loss function\n",
    "optimizer = optim.SGD(model.parameters(),\n",
    "                    lr=1e-3,\n",
    "                    momentum=0.9,\n",
    "#                       Momentum is a technique that accelerates gradient descent by accumulating a “velocity” vector in parameter space.\n",
    "# Instead of just moving in the direction of the current gradient, momentum remembers past gradients to smooth the updates.\n",
    "                    weight_decay=1e-4\n",
    "                     # weight decay Encourages the network to learn simpler, smoother mappings.\n",
    "                     ) #\n",
    "loss_fn = nn.CrossEntropyLoss()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "31a4f59c-99e6-4081-a1e4-ae4916634877",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 1 batches: 100%|██████████| 320/320 [01:50<00:00,  2.88it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2025-08-24 22:42:38.265523 Epoch 1, Training loss 0.2633039543405175, Validation loss 0.15964662106707692\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 2 batches: 100%|██████████| 320/320 [01:46<00:00,  3.02it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2025-08-24 22:44:32.191070 Epoch 2, Training loss 0.15434061667183413, Validation loss 0.19231202062219382\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 3 batches: 100%|██████████| 320/320 [01:49<00:00,  2.92it/s]\n",
      "Epoch 4 batches: 100%|██████████| 320/320 [01:47<00:00,  2.97it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2025-08-24 22:48:16.203702 Epoch 4, Training loss 0.12478437168174424, Validation loss 0.10536358784884214\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 5 batches: 100%|██████████| 320/320 [01:45<00:00,  3.05it/s]\n",
      "Epoch 6 batches: 100%|██████████| 320/320 [01:56<00:00,  2.74it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2025-08-24 22:52:04.059874 Epoch 6, Training loss 0.11163697743031661, Validation loss 0.09136132695712149\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 7 batches: 100%|██████████| 320/320 [01:38<00:00,  3.25it/s]\n",
      "Epoch 8 batches: 100%|██████████| 320/320 [01:46<00:00,  3.01it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2025-08-24 22:55:35.479441 Epoch 8, Training loss 0.09899017311399802, Validation loss 0.07920753569342195\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 9 batches: 100%|██████████| 320/320 [01:46<00:00,  3.00it/s]\n",
      "Epoch 10 batches: 100%|██████████| 320/320 [01:27<00:00,  3.65it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2025-08-24 22:58:56.311841 Epoch 10, Training loss 0.0951527620840352, Validation loss 0.11094426698982715\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 11 batches: 100%|██████████| 320/320 [01:32<00:00,  3.46it/s]\n",
      "Epoch 12 batches: 100%|██████████| 320/320 [01:27<00:00,  3.64it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2025-08-24 23:02:02.987800 Epoch 12, Training loss 0.08858228508324828, Validation loss 0.06188474288210273\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 13 batches: 100%|██████████| 320/320 [01:28<00:00,  3.63it/s]\n",
      "Epoch 14 batches: 100%|██████████| 320/320 [01:28<00:00,  3.63it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2025-08-24 23:05:05.414438 Epoch 14, Training loss 0.08608202228788286, Validation loss 0.06566031230613589\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 15 batches: 100%|██████████| 320/320 [01:49<00:00,  2.91it/s]\n",
      "Epoch 16 batches: 100%|██████████| 320/320 [01:27<00:00,  3.67it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2025-08-24 23:08:28.403245 Epoch 16, Training loss 0.08190540621289984, Validation loss 0.06167816421948373\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 17 batches: 100%|██████████| 320/320 [01:44<00:00,  3.05it/s]\n",
      "Epoch 18 batches: 100%|██████████| 320/320 [01:27<00:00,  3.66it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2025-08-24 23:11:47.312306 Epoch 18, Training loss 0.07801481284986948, Validation loss 0.0635686487541534\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 19 batches: 100%|██████████| 320/320 [01:47<00:00,  2.99it/s]\n",
      "Epoch 20 batches: 100%|██████████| 320/320 [01:20<00:00,  3.98it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2025-08-24 23:15:00.573301 Epoch 20, Training loss 0.07381812089588494, Validation loss 0.11788077496457845\n"
     ]
    }
   ],
   "source": [
    "# execution\n",
    "h, model = training_loop(\n",
    "    n_epochs = 20,\n",
    "    optimizer = optimizer,\n",
    "    model = model,\n",
    "    loss_fn = loss_fn,\n",
    "    train_loader = train_loader, \n",
    "    val_loader = val_loader,\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "c9507b2b-4c28-44cc-997a-ca8acc7a56fe",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Save only the state_dict (recommended)\n",
    "torch.save(model.state_dict(), \"model.pth\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "05b49990-963c-48fb-90b1-92c15d08c7aa",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy train: 0.9573\n",
      "Accuracy val: 0.9718\n"
     ]
    }
   ],
   "source": [
    "def validate(model, train_loader, val_loader):\n",
    "    model.eval()  # set to eval mode\n",
    "    misclassified = {\"train\": [], \"val\": []}  # store tuples (image, true_label, pred_label)\n",
    "    classified = {\"train\": [], \"val\": []} \n",
    "    \n",
    "    for name, loader in [(\"train\", train_loader), (\"val\", val_loader)]:\n",
    "        correct = 0\n",
    "        total = 0\n",
    "        with torch.no_grad():\n",
    "            for imgs, labels in loader:\n",
    "                imgs = imgs.to(device)\n",
    "                labels = labels.to(device)\n",
    "            \n",
    "                outputs = model(imgs)\n",
    "                _, predicted = torch.max(outputs, dim=1)\n",
    "                total += labels.size(0)\n",
    "                correct += (predicted == labels).sum().item()\n",
    "                \n",
    "                # Find misclassified samples indices in the batch\n",
    "                mis_mask = (predicted != labels)\n",
    "                if mis_mask.any():\n",
    "                    mis_imgs = imgs[mis_mask].cpu()  # bring to CPU for easy processing\n",
    "                    mis_true = labels[mis_mask].cpu()\n",
    "                    mis_pred = predicted[mis_mask].cpu()\n",
    "                    # Store as tuples for later use\n",
    "                    misclassified[name].extend(zip(mis_imgs, mis_true, mis_pred))\n",
    "\n",
    "                # Find misclassified samples indices in the batch\n",
    "                clas_mask = (predicted == labels)\n",
    "                if clas_mask.any():\n",
    "                    clas_imgs = imgs[clas_mask].cpu()  # bring to CPU for easy processing\n",
    "                    clas_true = labels[clas_mask].cpu()\n",
    "                    clas_pred = predicted[clas_mask].cpu()\n",
    "                    # Store as tuples for later use\n",
    "                    classified[name].extend(zip(clas_imgs, clas_true, clas_pred))\n",
    "        \n",
    "        print(f\"Accuracy {name}: {correct / total:.4f}\")\n",
    "\n",
    "    return classified, misclassified\n",
    "\n",
    "classified, misclassified = validate(model, train_loader, val_loader)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "f3adb655-1239-44fd-91d9-b1b8eef18188",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Label names: ['field', 'forest', 'sea', 'urban']\n",
      "Index to Label: {0: 'field', 1: 'forest', 2: 'sea', 3: 'urban'}\n",
      "Label to Index: {'field': 0, 'forest': 1, 'sea': 2, 'urban': 3}\n"
     ]
    }
   ],
   "source": [
    "# Look at features (they describe label mappings)\n",
    "features = dataset[\"train\"].features\n",
    "\n",
    "# Extract label names\n",
    "label_names = features[\"label\"].names\n",
    "print(\"Label names:\", label_names)\n",
    "\n",
    "# Create mapping\n",
    "idx2label = {i: label for i, label in enumerate(label_names)}\n",
    "label2idx = {label: i for i, label in enumerate(label_names)}\n",
    "\n",
    "print(\"Index to Label:\", idx2label)\n",
    "print(\"Label to Index:\", label2idx)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "fe193078-b88a-40b6-af55-a9af141a509c",
   "metadata": {},
   "outputs": [],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "\n",
    "history = h\n",
    "\n",
    "epochs = list(range(1, len(history[\"train_loss\"]) + 1))\n",
    "\n",
    "# Extract only the epochs and values where validation loss is not None\n",
    "val_epochs = [i+1 for i, v in enumerate(history[\"validation_loss\"]) if v is not None]\n",
    "val_values = [v for v in history[\"validation_loss\"] if v is not None]\n",
    "\n",
    "plt.plot(epochs, history[\"train_loss\"], marker='o', label=\"Training Loss\")\n",
    "\n",
    "# Line connecting only existing validation points\n",
    "plt.plot(val_epochs, val_values, label=\"Validation Loss\", color='orange')\n",
    "\n",
    "# Markers for the actual validation points\n",
    "plt.scatter(val_epochs, val_values, marker='s', color='orange')\n",
    "\n",
    "plt.xlabel(\"Epoch\")\n",
    "plt.ylabel(\"Loss\")\n",
    "plt.legend()\n",
    "plt.grid(True)\n",
    "plt.savefig(\"images/validation_training_loss.png\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "58bcf351-d236-4269-a043-787f0b643fbc",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "sea 2 forest\n",
      "sea 2 sea\n",
      "sea 2 field\n",
      "urban 3 urban\n",
      "field 0 field\n",
      "sea 2 sea\n"
     ]
    }
   ],
   "source": [
    "from PIL import Image\n",
    "\n",
    "img1 = Image.open(\"./images/mypix1.png\").convert(\"RGB\") #forest\n",
    "img2 = Image.open(\"./images/mypix2.png\").convert(\"RGB\") #sea\n",
    "img3 = Image.open(\"./images/mypix3.png\").convert(\"RGB\") #field\n",
    "img4 = Image.open(\"./images/mypix4.png\").convert(\"RGB\") #urban\n",
    "img5 = Image.open(\"./images/mypix5.png\").convert(\"RGB\") #field\n",
    "img6 = Image.open(\"./images/mypix6.png\").convert(\"RGB\") #sea\n",
    "\n",
    "img_list = [\n",
    "    img1, img2, img3, img4, img5, img6, \n",
    "]\n",
    "category_list = [\n",
    "    \"forest\", \"sea\", \"field\", \"urban\", \"field\", \"sea\", \n",
    "]\n",
    "\n",
    "# Define the same transforms used in your model\n",
    "transform = transforms.Compose([\n",
    "    transforms.Resize((256, 256)),       # match your model input size\n",
    "    transforms.ToTensor(),               # convert to [C,H,W] tensor\n",
    "    transforms.Normalize(\n",
    "        mean=[0.2918, 0.3070, 0.2985],           # same as during training\n",
    "        std=[0.1065, 0.0985, 0.0877]\n",
    "    )\n",
    "])\n",
    "\n",
    "dict_labels = {0: 'field', 1: 'forest', 2: 'sea', 3: 'urban'}\n",
    "\n",
    "for i, c in zip(img_list, category_list):\n",
    "\n",
    "    # Apply transforms\n",
    "    img_tensor = transform(i).unsqueeze(0)  # add batch dimension [1, C, H, W]\n",
    "    \n",
    "    img_tensor = img_tensor.to(device)\n",
    "    \n",
    "    outputs = model(img_tensor)\n",
    "    _, predicted = torch.max(outputs, dim=1)\n",
    "    print(dict_labels[int(predicted)], int(predicted), c)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "41a17bad-2b6b-4e0a-98af-c96b22e9ff30",
   "metadata": {},
   "outputs": [],
   "source": [
    "# import matplotlib.pyplot as plt\n",
    "# import math\n",
    "\n",
    "# def plot_images(misclassified, classes, title=None, n_images=20, subset=\"val\"):\n",
    "#     imgs = misclassified[subset][:n_images]\n",
    "#     n_cols = 5\n",
    "#     n_rows = math.ceil(len(imgs) / n_cols)\n",
    "    \n",
    "#     fig, axes = plt.subplots(n_rows, n_cols, figsize=(3 * n_cols, 3 * n_rows))\n",
    "#     axes = axes.flatten()  # flatten in case of multiple rows\n",
    "    \n",
    "#     for ax in axes[len(imgs):]:\n",
    "#         ax.axis('off')  # turn off extra axes\n",
    "    \n",
    "#     for i, (img, true_lbl, pred_lbl) in enumerate(imgs):\n",
    "#         img_np = img.permute(1, 2, 0).numpy()\n",
    "#         img_np = (img_np - img_np.min()) / (img_np.max() - img_np.min())  # normalize 0-1\n",
    "        \n",
    "#         axes[i].imshow(img_np)\n",
    "#         axes[i].set_title(f\"True: {classes[true_lbl.item()]}\\nPred: {classes[pred_lbl.item()]}\")\n",
    "#         axes[i].axis('off')\n",
    "    \n",
    "#     plt.tight_layout()\n",
    "#     if title is not None:\n",
    "#         plt.suptitle(title, fontsize=16)\n",
    "#         plt.tight_layout(rect=[0, 0, 1, 0.97])\n",
    "#         plt.savefig(title.replace(\" \", \"_\") + \".png\")\n",
    "\n",
    "#     plt.show()\n",
    "\n",
    "# plot_images(misclassified, classes=idx2label,\n",
    "#             title = \"Misclassified pictures\",\n",
    "#             n_images=40, subset=\"val\")\n",
    "\n",
    "# plot_images(classified, classes=idx2label,\n",
    "#             title = \"Well-classified pictures\",\n",
    "#             n_images=40, subset=\"val\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "02fa86cb-5a81-4de5-83f3-c552f17801e5",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.13.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}