facial-expression-recognition.ipynb

{
 "cells": [
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Prepoznavanje facijalnih ekspresija (Facial expression recognition)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import torch\n",
    "from torch import nn\n",
    "from torch.utils.data import Dataset, DataLoader\n",
    "from torchvision import transforms\n",
    "from pytorchtools import EarlyStopping\n",
    "from sklearn.preprocessing import LabelBinarizer\n",
    "from sklearn.metrics import confusion_matrix\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Skup podataka (Dataset)\n",
    "\n",
    "Izabrani skup podataka za obradu je skup facijalnih ekspresija pod nazivom **FER2013**.\n",
    "\n",
    "Podaci su javno dostupni i preuzeti sa web stranice *Kaggle*.\n",
    "\n",
    "Cilj obrade podataka je obučavanje modela za klasifikaciju jedne facijalne ekspresije prikazane na slici. \n",
    "\n",
    "Skup podataka sadrži skoro od 36.000 slika, koje predstavljaju 7 klasa facijalnih ekspresija."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "class FacialExpressionDataset(Dataset):\n",
    "    def __init__(self, path, type, transform):\n",
    "        super().__init__()\n",
    "\n",
    "        self.features = []\n",
    "        self.labels = []\n",
    "\n",
    "        self.type = type\n",
    "        self.transform = transform\n",
    "        self.label_encoder = LabelBinarizer()\n",
    "        self._read(path)\n",
    "\n",
    "    def _read(self, path):\n",
    "        data = pd.read_csv(path)\n",
    "        for index, row in data.iterrows():\n",
    "            if self.type == row['Usage']:\n",
    "                self.features.append(self._load_image(row['pixels']))\n",
    "                self.labels.append(row['emotion'])\n",
    "        self.labels = self.label_encoder.fit_transform(self.labels).astype(np.float32)\n",
    "\n",
    "    def _load_image(self, pixels):\n",
    "        pixels = np.array(list(map(int, pixels.split())), dtype=np.float32)/255\n",
    "        return self.transform(pixels.reshape((48, 48, 1)))\n",
    "\n",
    "    def __len__(self):\n",
    "        return len(self.features)\n",
    "\n",
    "    def __getitem__(self, index):\n",
    "        return self.features[index], self.labels[index]"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Pripremna obrada podataka\n",
    "\n",
    "Svaka slika je konvertovana u tenzor."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "transform = transforms.Compose([\n",
    "    transforms.ToTensor()\n",
    "])\n",
    "\n",
    "dataset_path = \"dataset/fer2013.csv\"\n",
    "\n",
    "train_dataset = FacialExpressionDataset(dataset_path, \"Training\", transform)\n",
    "valid_dataset = FacialExpressionDataset(dataset_path, \"PublicTest\", transform)\n",
    "test_dataset = FacialExpressionDataset(dataset_path, \"PrivateTest\", transform)"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Informacije o broju uzoraka i klasa\n",
    "\n",
    "Podaci su već podeljeni na trening, validacioni i test skup, a njihov odnos je takav da je 80% podataka za trening, a po 10% za validaciju i testiranje."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "classes = {\n",
    "    0: 'Angry', \n",
    "    1: 'Disgust', \n",
    "    2: 'Fear', \n",
    "    3: 'Happy', \n",
    "    4: 'Sad', \n",
    "    5: 'Suprise', \n",
    "    6: 'Neutral'\n",
    "}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Ukupan broj uzoraka = 35887\n",
      "Broj uzoraka za trening = 28709 (80.00%)\n",
      "Broj uzoraka za validaciju = 3589 (10.00%)\n",
      "Broj uzoraka za testiranje = 3589 (10.00%)\n",
      "Broj klasa = 7\n"
     ]
    }
   ],
   "source": [
    "total_examples = len(train_dataset) + len(valid_dataset) + len(test_dataset)\n",
    "\n",
    "print(f\"Ukupan broj uzoraka = {total_examples}\")\n",
    "print(f\"Broj uzoraka za trening = {len(train_dataset)} ({len(train_dataset)/total_examples*100:.2f}%)\")\n",
    "print(f\"Broj uzoraka za validaciju = {len(valid_dataset)} ({len(valid_dataset)/total_examples*100:.2f}%)\")\n",
    "print(f\"Broj uzoraka za testiranje = {len(test_dataset)} ({len(test_dataset)/total_examples*100:.2f}%)\")\n",
    "print(f\"Broj klasa = {len(classes)}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 2400x400 with 3 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "def setup_axe(axe, data, encoder, title):\n",
    "    display_data = encoder.inverse_transform(data)\n",
    "    occurrences = [np.count_nonzero(display_data == key) for key in classes.keys()]\n",
    "    axe.bar(classes.values(), occurrences)\n",
    "    axe.set_xlabel(\"Emotions\")\n",
    "    axe.set_ylabel(\"Number\")\n",
    "    axe.set_title(title)\n",
    "\n",
    "fig, axes = plt.subplots(1, 3, figsize=(24,4), sharey=True)\n",
    "setup_axe(axes[0], train_dataset.labels, train_dataset.label_encoder, 'train')\n",
    "setup_axe(axes[1], valid_dataset.labels, valid_dataset.label_encoder, 'valid')\n",
    "setup_axe(axes[2], test_dataset.labels, test_dataset.label_encoder, 'test')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Nasumični predstavnici svih klasa u skupu podataka\n",
    "\n",
    "Kao što se može videti iz primera klasa, skup podataka je izazovan iz razloga sto je veliki broj slika mutan, a neke slike su ili previše ili nedovoljno osvetljene."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 2400x200 with 7 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, axes = plt.subplots(1, 7, figsize=(24, 2))\n",
    "display_labels = test_dataset.label_encoder.inverse_transform(test_dataset.labels)\n",
    "\n",
    "for class_id, class_name in classes.items():\n",
    "    axe = axes[class_id]\n",
    "    axe.set_title(class_name)\n",
    "    axe.set_axis_off()\n",
    "    i = np.random.choice(np.where(display_labels == class_id)[0])\n",
    "    image = np.transpose(test_dataset[i][0], (1, 2, 0))\n",
    "    axe.imshow(image, cmap=\"gray\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Određivanje hardvera za računanja\n",
    "\n",
    "Ukoliko je CUDA dostupan, koristiće se GPU, u suprotnom koristi se CPU."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Hardver koji ce se upotrebljivati = cuda\n"
     ]
    }
   ],
   "source": [
    "if torch.cuda.is_available():\n",
    "    device = torch.device(\"cuda\")\n",
    "    torch.cuda.empty_cache()\n",
    "else:\n",
    "    device = torch.device(\"cpu\")\n",
    "\n",
    "print(f\"Hardver koji ce se upotrebljivati = {device}\")\n",
    "\n",
    "def to_device(x, y):\n",
    "    return x.to(device), y.to(device)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "class WrappedDataLoader:\n",
    "    def __init__(self, dl, func):\n",
    "        self.dl = dl\n",
    "        self.func = func\n",
    "\n",
    "    def __len__(self):\n",
    "        return len(self.dl)\n",
    "\n",
    "    def __iter__(self):\n",
    "        batches = iter(self.dl)\n",
    "        for b in batches:\n",
    "            yield (self.func(*b))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_dl = WrappedDataLoader(DataLoader(train_dataset, batch_size=64, shuffle=True), to_device) \n",
    "valid_dl = WrappedDataLoader(DataLoader(valid_dataset, batch_size=64, shuffle=False), to_device) \n",
    "test_dl = WrappedDataLoader(DataLoader(test_dataset, batch_size=64, shuffle=False), to_device)"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Konvolucijska neuronska mreža (Convolutional Neural Network)\n",
    "\n",
    "Konvolucijska neuronska mreza (CNN) je vrsta feedforward veštacke neuronske mreže koja se često primenjuje za analizu slika. Arhitektura ove mreže sačinjena iz tri osnovne komponente:\n",
    "\n",
    "##### 1. Konvolucioni sloj (Convolution Layer)\n",
    "\n",
    "Konvulacija je matematički operator koji od dve funkcije, $f$ i $g$, proizvodi treću funkciju $f*g$ koja izražava koliko je oblik jedne funkcije modifikovan drugom.\n",
    "\n",
    "Ovo je najvažniji sloj u okviru konvolucijskih mreža, to jest na njemu je bazirana mreža.\n",
    "\n",
    "U konvulucionom sloju koristi se kernel, to jest filter, koji izvlači relevantne osobine iz ulaza upotrebom operacije konvulacije. Oni se koriste za izoštravanje slike i zamućenje slika, kao i za detektovanje ivica.\n",
    "\n",
    "##### 2. Sloj sažimanja (Pooling Layer)\n",
    "\n",
    "Sloj sažimanja je sloj koji se periodično javlja, a čija je uloga smanjivanje dimenzija slike kako bi se smanjio broj parametara.\n",
    "\n",
    "Postoje dve opcije koje se mogu izabrati u okviru ovog sloja: sažimanje maksimumom (Max Pooling) i sažimanje pomoću srednje vrednosti (Average Pooling).\n",
    "\n",
    "##### 3. Potpuno povezan sloj (Fully Connected Layer)\n",
    "\n",
    "Zbog velikog broj parametra, neuroni u prethodnom delu mreže ne moraju biti potpuno povezani, te se zbog toga u završnim slojevima mreže pojavljuju potpuno povezani slojevi. \n",
    "\n",
    "To nije pravilo, ali se u najvećem broju slučajeva realizuje na taj način, jer su dimenzije na ulazu u ovaj sloj znatno manje."
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Aktivaciona funkcija\n",
    "\n",
    "Aktivaciona funkcija se koristi za poboljšanje perfomansi mreže; za aktivacionu funkciju ovde je iskorišćena ReLU funkcija koja se veoma cešto koristi za skrivene slojeve u neuronskim mrežama."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "class CNN(nn.Module):\n",
    "    def __init__(self, classes):\n",
    "        super(CNN, self).__init__()\n",
    "        self.conv1 = nn.Sequential(\n",
    "            nn.Conv2d(1, 256, kernel_size=3, padding=1),\n",
    "            nn.BatchNorm2d(256),\n",
    "            nn.ReLU(),\n",
    "        )\n",
    "        self.conv2 = nn.Sequential(\n",
    "            nn.Conv2d(256, 256, kernel_size=3, padding=1),\n",
    "            nn.BatchNorm2d(256),\n",
    "            nn.ReLU(),\n",
    "            nn.MaxPool2d(kernel_size=2, stride=2)\n",
    "        )\n",
    "        self.conv3 = nn.Sequential(\n",
    "            nn.Conv2d(256, 128, kernel_size=3, padding=1),\n",
    "            nn.BatchNorm2d(128),\n",
    "            nn.ReLU()\n",
    "        )\n",
    "        self.conv4 = nn.Sequential(\n",
    "            nn.Conv2d(128, 128, kernel_size=3, padding=1),\n",
    "            nn.BatchNorm2d(128),\n",
    "            nn.ReLU(),\n",
    "            nn.MaxPool2d(kernel_size=2, stride=2)\n",
    "        )\n",
    "        self.conv5 = nn.Sequential(\n",
    "            nn.Conv2d(128, 64, kernel_size=3, padding=1),\n",
    "            nn.BatchNorm2d(64),\n",
    "            nn.ReLU(),\n",
    "        )\n",
    "        self.conv6 = nn.Sequential(\n",
    "            nn.Conv2d(64, 64, kernel_size=3, padding=1),\n",
    "            nn.BatchNorm2d(64),\n",
    "            nn.ReLU(),\n",
    "            nn.MaxPool2d(kernel_size=2, stride=2)\n",
    "        )\n",
    "        self.fc1 = nn.Sequential(\n",
    "            nn.Linear(2304, 512),\n",
    "            nn.BatchNorm1d(512),\n",
    "            nn.ReLU()\n",
    "        )\n",
    "        self.fc2 = nn.Sequential(\n",
    "            nn.Linear(512, 256),\n",
    "            nn.BatchNorm1d(256),\n",
    "            nn.ReLU()\n",
    "        )\n",
    "        self.fc3 = nn.Sequential(\n",
    "            nn.Linear(256, 128),\n",
    "            nn.BatchNorm1d(128),\n",
    "            nn.ReLU()\n",
    "        )\n",
    "        self.out = nn.Linear(128, classes)\n",
    "\n",
    "    def forward(self, x):\n",
    "        x = self.conv1(x)\n",
    "        x = self.conv2(x)\n",
    "        x = self.conv3(x)\n",
    "        x = self.conv4(x)\n",
    "        x = self.conv5(x)\n",
    "        x = self.conv6(x)\n",
    "        x = x.view(-1, 2304)\n",
    "        x = self.fc1(x)\n",
    "        x = self.fc2(x)\n",
    "        x = self.fc3(x)\n",
    "        return self.out(x)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model ima 2494855 parametara koji treba da se treniraju\n"
     ]
    }
   ],
   "source": [
    "def count_parameters(model):\n",
    "    return sum(p.numel() for p in model.parameters() if p.requires_grad)\n",
    "\n",
    "print(f\"Model ima {count_parameters(CNN(len(classes)))} parametara koji treba da se treniraju\")"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Definisanje funkcija za treniranje i testiranje.\n",
    "\n",
    "U metodi za treniranje koristi se objekat klase EarlyStopping koji će pratiti gubitak validacije.\n",
    "\n",
    "Ukoliko nakon određene toleranicije ona ne nastavi da se smanjuje - obustaviće treniranje i vratiti model na parametere kada je prvi put dostignuta najmanja grešku kako bi se sprečio overfitting."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "def evaluate(model, loss_func, dl):\n",
    "    loss = 0\n",
    "    total = 0\n",
    "    correct = 0\n",
    "\n",
    "    with torch.no_grad():\n",
    "        for fv, y in dl:\n",
    "            yp = model.forward(fv)\n",
    "            loss += loss_func(yp, y)\n",
    "            max_index = torch.argmax(yp, dim=1).cpu()\n",
    "            total += yp.size(0)\n",
    "            y_transformed = torch.tensor(list(map(lambda y: np.argmax(y), y.cpu().numpy())))\n",
    "            correct += (y_transformed == max_index).sum().item()\n",
    "\n",
    "    return loss/len(dl), correct/total"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "def train(epochs, patience, model, loss_func, optimizer, train_dl, valid_dl):\n",
    "    train_losses = []\n",
    "    valid_losses = []\n",
    "\n",
    "    early_stopping = EarlyStopping(patience=patience, path=\"models/checkpoint.pth\")\n",
    "\n",
    "    for epoch in range(epochs):\n",
    "        total_train_loss = 0\n",
    "\n",
    "        for fv, y in train_dl:\n",
    "            optimizer.zero_grad()\n",
    "            yp = model.forward(fv)\n",
    "            loss = loss_func(yp, y)\n",
    "            loss.backward()\n",
    "            total_train_loss += loss\n",
    "            optimizer.step()\n",
    "        \n",
    "        train_loss = total_train_loss / len(train_dl)\n",
    "        valid_loss, valid_accuracy = evaluate(model, loss_func, valid_dl)\n",
    "\n",
    "        train_losses.append(train_loss)\n",
    "        valid_losses.append(valid_loss)\n",
    "\n",
    "        print(f\"[Epoch {epoch+1}/{epochs}] Training loss: {train_loss:.2f}\\t\"\n",
    "                f\"Validation loss: {valid_loss:.2f}\\tValidation accuracy: {valid_accuracy * 100:.2f}%\")\n",
    "\n",
    "        early_stopping(valid_loss, model)\n",
    "        if early_stopping.early_stop:\n",
    "            print(\"Early stopping\")\n",
    "            break\n",
    "\n",
    "    model.load_state_dict(torch.load(\"models/checkpoint.pth\"))\n",
    "\n",
    "    return model, train_losses, valid_losses"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "def test(model, loss_func, test_dl):\n",
    "    test_loss, test_accuracy = evaluate(model, loss_func, test_dl)\n",
    "    print(f\"Test loss: {test_loss:.2f}\\tTest accuracy: {test_accuracy * 100:.2f}%\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Gubitak unakrsne entropije (Cross-Entropy Loss)\n",
    "\n",
    "Gubitak unakrsne entropije određuje perfomanse modela za klasifikaciju čiji su izlazi vrednosti verovatnoće između 0 i 1.\n",
    "\n",
    "Ova vrednost se povećava kada predviđena verovatnoća divergira od stvarne labele.\n",
    "\n",
    "U slučaju PyTorch biblioteke, klasa CrossEntropyLoss očekuje sirove podatke, to jest ne treba ih propustiti kroz softmax funkciju, jer ona interno primenjuje softmax."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Upotreba Adam (Adaptive Moment Estimation) algoritma za optimizaciju\n",
    "\n",
    "Jedan od najpopularnijih i najčešće upotrebljavanih algoritama za mašinsko učenje je upravo SGD, i on ima brojna proširenja i varijacije.\n",
    "\n",
    "Adam algoritam je proširenje SGD algoritma, to jest kombinacija RMSProp (Root Mean Squared Propagation) i SGD sa momentumom.\n",
    "\n",
    "U poslednje vreme ovaj algoritam ima veliku primenu u dubokom učenju za kompjutersku viziju i obradu prirodnih jezika."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "net = CNN(len(classes)).to(device)\n",
    "loss_function = nn.CrossEntropyLoss().to(device)\n",
    "optimizer = torch.optim.Adam(net.parameters(), lr=0.001)\n",
    "epochs = 20\n",
    "patience = 5"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[Epoch 1/20] Training loss: 1.48\tValidation loss: 1.29\tValidation accuracy: 50.93%\n",
      "[Epoch 2/20] Training loss: 1.18\tValidation loss: 1.17\tValidation accuracy: 55.31%\n",
      "[Epoch 3/20] Training loss: 1.04\tValidation loss: 1.12\tValidation accuracy: 58.54%\n",
      "[Epoch 4/20] Training loss: 0.93\tValidation loss: 1.13\tValidation accuracy: 58.54%\n",
      "[Epoch 5/20] Training loss: 0.82\tValidation loss: 1.10\tValidation accuracy: 60.27%\n",
      "[Epoch 6/20] Training loss: 0.69\tValidation loss: 1.14\tValidation accuracy: 61.35%\n",
      "[Epoch 7/20] Training loss: 0.56\tValidation loss: 1.23\tValidation accuracy: 61.10%\n",
      "[Epoch 8/20] Training loss: 0.43\tValidation loss: 1.36\tValidation accuracy: 60.82%\n",
      "[Epoch 9/20] Training loss: 0.32\tValidation loss: 1.52\tValidation accuracy: 60.02%\n",
      "[Epoch 10/20] Training loss: 0.25\tValidation loss: 1.72\tValidation accuracy: 59.82%\n",
      "Early stopping\n",
      "Test loss: 1.08\tTest accuracy: 60.94%\n"
     ]
    }
   ],
   "source": [
    "net, train_losses, valid_losses = train(epochs, patience, net, loss_function, optimizer, train_dl, valid_dl)\n",
    "test(net, loss_function, test_dl)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 800x400 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "train_losses_ = [x.cpu().detach() for x in train_losses]\n",
    "valid_losses_ = [x.cpu().detach() for x in valid_losses]\n",
    "\n",
    "plt.figure(figsize=(8,4))\n",
    "plt.plot(range(1,len(train_losses_)+1),train_losses_, label='Training Loss')\n",
    "plt.plot(range(1,len(valid_losses_)+1),valid_losses_, label='Validation Loss')\n",
    "plt.axvline(valid_losses_.index(min(valid_losses_))+1, linestyle='--', color='r',label='Early Stopping Checkpoint')\n",
    "plt.xlabel('epochs')\n",
    "plt.ylabel('loss')\n",
    "plt.ylim(0, 2)\n",
    "plt.xlim(0, len(train_losses_)+1)\n",
    "plt.grid(True)\n",
    "plt.legend()\n",
    "plt.tight_layout()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "torch.save(net, \"models/net.pth\")"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Analiza rezultata\n",
    "\n",
    "Dobijena preciznost od 61% na test podacima je dobar rezultat s obzirom da je na takmičenju sa ovim skupom podataka najbolji rezultat bio 71.1%.\n",
    "\n",
    "Ovaj rezultat ide u prilog činjenici da je CNN danas favorit za probleme kompjuterske vizije."
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Matrica konfuzije (Confusion Matrix)\n",
    "\n",
    "Matrica konfuzije je tehnika koja se koristi za sumarizaciju perfomansi klasifikacionih algoritama.\n",
    "\n",
    "Kroz njenu upotrebu je lakše uočiti gde model daje dobre predikcije i kakve greške pravi.\n",
    "\n",
    "U ovoj matrici redovima su predstavljene stvarne vrednosti, a kolonama predikcije."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1000x600 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "labels = test_dataset.label_encoder.inverse_transform(test_dataset.labels)\n",
    "predictions = []\n",
    "with torch.no_grad():\n",
    "    for inputs, _ in test_dl: \n",
    "        predictions.extend(torch.argmax(net(inputs), dim=1).data.cpu().numpy())\n",
    "\n",
    "cf_matrix = confusion_matrix(labels, predictions)\n",
    "\n",
    "df_cm = pd.DataFrame(\n",
    "    data = [row / np.sum(row) for row in cf_matrix],\n",
    "    index = [i for i in classes.values()], \n",
    "    columns = [i for i in classes.values()]\n",
    ").round(2)\n",
    "\n",
    "plt.figure(figsize = (10,6))\n",
    "hm = sns.heatmap(df_cm, annot=True, cmap=\"Blues\")\n",
    "hm.set(xlabel=\"Predicted label\", ylabel=\"True label\");"
   ]
  }
 ],
 "metadata": {
  "interpreter": {
   "hash": "7f21d13ffba6745f94cd61defb1b93ca446141575ea451a7fcf879ce80ac9139"
  },
  "kernelspec": {
   "display_name": "Python 3.8.10 64-bit ('ai': venv)",
   "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": "3.10.6"
  },
  "orig_nbformat": 4
 },
 "nbformat": 4,
 "nbformat_minor": 2
}