arabic_letter_classification.ipynb

{
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "hh72Wfww9yt2"
      },
      "source": [
        "# Import"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "xRSoWJtK9yt3"
      },
      "outputs": [],
      "source": [
        "import matplotlib.pyplot as plt\n",
        "import pandas as pd\n",
        "import numpy as np\n",
        "import torch\n",
        "import torch.nn as nn\n",
        "from torch.utils.data import DataLoader, TensorDataset\n",
        "from torchsummary import summary\n",
        "from torch.optim import Adam\n",
        "from sklearn.metrics import confusion_matrix, f1_score, classification_report\n",
        "import seaborn as sn\n",
        "from tqdm import tqdm"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "np.set_printoptions(precision=4)"
      ],
      "metadata": {
        "id": "Kk-fRMZpXi-c"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')\n",
        "device"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "BH9qQbeOKy8T",
        "outputId": "2066ec3a-84f0-4f43-cf67-ba42e8d3ae93"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "device(type='cuda')"
            ]
          },
          "metadata": {},
          "execution_count": 2
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "SqQ43o7u9yt-"
      },
      "source": [
        "# Read Data"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Constants\n",
        "IMAGE_SIZE = (32, 32)\n",
        "NUM_LETTERS = 28"
      ],
      "metadata": {
        "id": "r0KHNB6Q6ftq"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "!unzip \"data.zip\""
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "xHmiifrjTDnC",
        "outputId": "2264c616-29ba-4688-b6b2-55c7ac15e7b3"
      },
      "execution_count": 2,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Archive:  data.zip\n",
            "  inflating: TrainImgs.csv           \n",
            "  inflating: TestImgs.csv            \n",
            "  inflating: TestLabels.csv          \n",
            "  inflating: TrainLabels.csv         \n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "dKkkCd059yuA",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "fe52c87e-caf5-4829-d008-fdebd891814b"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Shape of test images: (3360, 1024)\n",
            "Shape of test labels: (3360, 1)\n",
            "Shape of train images: (13440, 1024)\n",
            "Shape of train labels: (13440, 1)\n"
          ]
        }
      ],
      "source": [
        "train_images = pd.read_csv(\"TrainImgs.csv\", header=None)\n",
        "train_labels = pd.read_csv(\"TrainLabels.csv\", header=None)\n",
        "\n",
        "test_images = pd.read_csv(\"TestImgs.csv\", header=None)\n",
        "test_labels = pd.read_csv(\"TestLabels.csv\", header=None)\n",
        "\n",
        "print(f'Shape of test images: {test_images.shape}')\n",
        "print(f'Shape of test labels: {test_labels.shape}')\n",
        "print(f'Shape of train images: {train_images.shape}')\n",
        "print(f'Shape of train labels: {train_labels.shape}')"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "X_train = train_images.to_numpy().reshape((-1,1, *IMAGE_SIZE))\n",
        "X_test = test_images.to_numpy().reshape((-1,1, *IMAGE_SIZE))\n",
        "y_train = train_labels.to_numpy().squeeze() - 1 # Original labels are from 1 to 28\n",
        "y_test = test_labels.to_numpy().squeeze() - 1"
      ],
      "metadata": {
        "id": "b_MuodRl7GwS"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "# EDA"
      ],
      "metadata": {
        "id": "WF9qOVjiNFZA"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "plt.figure(figsize=(20, 6))\n",
        "train_counts = [(y_train == i).sum() for i in range(NUM_LETTERS)]\n",
        "test_counts = [(y_test == i).sum() for i in range(NUM_LETTERS)]\n",
        "index = np.arange(NUM_LETTERS)\n",
        "\n",
        "bar_width = 0.35\n",
        "plt.bar(index - bar_width/2, train_counts, color='b', label='train', width=bar_width)\n",
        "plt.bar(index + bar_width/2, test_counts, color='r', label='test', width=bar_width)\n",
        "\n",
        "plt.tight_layout()\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 307
        },
        "id": "mlwn3AMR-C7Y",
        "outputId": "f02ce503-891b-481b-b314-ec9da90e3cad"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 1440x432 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 467
        },
        "id": "H81ksJny9yuC",
        "outputId": "269d51ba-fd7d-49c4-f2ee-0b3c2b59d15b"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 864x576 with 28 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ],
      "source": [
        "plt.figure(figsize=(12, 8))\n",
        "class_indices = [np for i in range(NUM_LETTERS)]\n",
        "for i in range(NUM_LETTERS):\n",
        "    index = np.random.choice(np.where(y_train == i)[0])\n",
        "    plt.subplot(4, 7, i + 1)\n",
        "    plt.imshow(X_train[index][0].T)\n",
        "    plt.xticks([])\n",
        "    plt.yticks([])\n",
        "    plt.xlabel(i)"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Datasets and DataLoaders"
      ],
      "metadata": {
        "id": "XU0bE-nyNLw6"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Normalization\n",
        "X_train = (X_train / 255).astype('float32')\n",
        "X_test = (X_test / 255).astype('float32')"
      ],
      "metadata": {
        "id": "83Dzmw5TEoy_"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "train_set = TensorDataset(torch.tensor(X_train), torch.tensor(y_train))\n",
        "test_set = TensorDataset(torch.tensor(X_test), torch.tensor(y_test))\n",
        "\n",
        "BATCH_SIZE = 32\n",
        "train_loader = DataLoader(train_set, batch_size=BATCH_SIZE, shuffle=True)\n",
        "test_loader = DataLoader(test_set, batch_size=len(test_set), shuffle=True)"
      ],
      "metadata": {
        "id": "V-CP-CnOR_6W"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Model\n",
        "For the model, I used [ResNet](https://en.wikipedia.org/wiki/Residual_neural_network). It's a powerful convolutional model which takes advantage of skip connections in order to get rid of the vanishing gradient problem in deep networks."
      ],
      "metadata": {
        "id": "jBticFNaNSn-"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "class ResidualClassifier(nn.Module):\n",
        "    def __init__(self):\n",
        "        super().__init__()\n",
        "        self.seq1 = nn.Sequential(\n",
        "            nn.Conv2d(in_channels=1, out_channels=16, kernel_size=3, padding=1),\n",
        "            nn.ReLU(),\n",
        "            nn.MaxPool2d(kernel_size=2, stride=2)\n",
        "        )\n",
        "\n",
        "        self.res1 = nn.Sequential(\n",
        "            nn.Conv2d(in_channels=16, out_channels=16, kernel_size=3, padding=1),\n",
        "            nn.ReLU(),\n",
        "            nn.Conv2d(in_channels=16, out_channels=16, kernel_size=3, padding=1),\n",
        "            nn.ReLU()\n",
        "        )\n",
        "\n",
        "        self.seq2 = nn.Sequential(\n",
        "            nn.MaxPool2d(kernel_size=2, stride=2),\n",
        "            nn.Conv2d(in_channels=16, out_channels=64, kernel_size=3, padding=1),\n",
        "            nn.ReLU()\n",
        "        )\n",
        "\n",
        "        self.res2 = nn.Sequential(\n",
        "            nn.Conv2d(in_channels=64, out_channels=64, kernel_size=3, padding=1),\n",
        "            nn.ReLU(),\n",
        "            nn.Conv2d(in_channels=64, out_channels=64, kernel_size=3, padding=1),\n",
        "            nn.ReLU()\n",
        "        )\n",
        "\n",
        "        self.seq3 = nn.Sequential(\n",
        "            nn.AdaptiveAvgPool2d(output_size=(1, 1)),\n",
        "            nn.Flatten(),\n",
        "            nn.Linear(in_features=64, out_features=NUM_LETTERS)\n",
        "        )\n",
        "\n",
        "\n",
        "    def forward(self, x):\n",
        "        y1 = self.seq1(x)\n",
        "        y2 = self.res1(y1) + y1\n",
        "        y3 = self.seq2(y2)\n",
        "        y4 = self.res2(y3) + y3\n",
        "        y5 = self.seq3(y4)\n",
        "        return y5"
      ],
      "metadata": {
        "id": "ln-3URtVIXtS"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Defining the model and other parameters"
      ],
      "metadata": {
        "id": "-MUQ4Df8N-Tf"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "\"torch.manual_seed(0)\n",
        "model = ResidualClassifier()\n",
        "model.to(device)\n",
        "criterion = nn.CrossEntropyLoss()\n",
        "optimizer = Adam(model.parameters(), lr = 0.01)\n",
        "scheduler = torch.optim.lr_scheduler.ExponentialLR(optimizer, gamma=0.97)\n",
        "\n",
        "summary(model, (1, 32, 32))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "M0njmNqzIrYY",
        "outputId": "47210415-a87a-4d48-c57d-25939d5f4627"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "----------------------------------------------------------------\n",
            "        Layer (type)               Output Shape         Param #\n",
            "================================================================\n",
            "            Conv2d-1           [-1, 16, 32, 32]             160\n",
            "              ReLU-2           [-1, 16, 32, 32]               0\n",
            "         MaxPool2d-3           [-1, 16, 16, 16]               0\n",
            "            Conv2d-4           [-1, 16, 16, 16]           2,320\n",
            "              ReLU-5           [-1, 16, 16, 16]               0\n",
            "            Conv2d-6           [-1, 16, 16, 16]           2,320\n",
            "              ReLU-7           [-1, 16, 16, 16]               0\n",
            "         MaxPool2d-8             [-1, 16, 8, 8]               0\n",
            "            Conv2d-9             [-1, 64, 8, 8]           9,280\n",
            "             ReLU-10             [-1, 64, 8, 8]               0\n",
            "           Conv2d-11             [-1, 64, 8, 8]          36,928\n",
            "             ReLU-12             [-1, 64, 8, 8]               0\n",
            "           Conv2d-13             [-1, 64, 8, 8]          36,928\n",
            "             ReLU-14             [-1, 64, 8, 8]               0\n",
            "AdaptiveAvgPool2d-15             [-1, 64, 1, 1]               0\n",
            "          Flatten-16                   [-1, 64]               0\n",
            "           Linear-17                   [-1, 28]           1,820\n",
            "================================================================\n",
            "Total params: 89,756\n",
            "Trainable params: 89,756\n",
            "Non-trainable params: 0\n",
            "----------------------------------------------------------------\n",
            "Input size (MB): 0.00\n",
            "Forward/backward pass size (MB): 0.60\n",
            "Params size (MB): 0.34\n",
            "Estimated Total Size (MB): 0.95\n",
            "----------------------------------------------------------------\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Training"
      ],
      "metadata": {
        "id": "HMEHd6AEOOCf"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "EPOCHS = 30\n",
        "\n",
        "metrics = {\n",
        "        'train loss': [], 'test loss': [],\n",
        "        'train acc': [], 'test acc': [],\n",
        "        'train f1': [], 'test f1': []\n",
        "    }\n",
        "for epoch in range(EPOCHS):\n",
        "    train_loss, test_loss = 0, 0\n",
        "    train_acc, test_acc = 0, 0\n",
        "    model.train(True)\n",
        "    all_preds = []\n",
        "    all_labels = []\n",
        "    for images, labels in train_loader:\n",
        "        all_labels += labels\n",
        "        images = images.to(device)\n",
        "        labels = labels.to(device)\n",
        "        optimizer.zero_grad()\n",
        "        outputs = model(images)\n",
        "        loss = criterion(outputs , labels)\n",
        "        loss.backward()\n",
        "        optimizer.step()\n",
        "\n",
        "        all_preds += torch.argmax(outputs, 1).to('cpu')\n",
        "        train_loss += loss\n",
        "    scheduler.step()\n",
        "    train_acc = sum([int(p == l) for p, l in zip(all_preds, all_labels)])/len(train_set)\n",
        "    train_f1 = f1_score(all_labels, all_preds, average='macro')\n",
        "    train_loss /= (len(train_set) // BATCH_SIZE + 1)\n",
        "\n",
        "    with torch.no_grad():\n",
        "        for images, labels in test_loader:\n",
        "            cpu_labels = labels\n",
        "            images = images.to(device)\n",
        "            labels = labels.to(device)\n",
        "            optimizer.zero_grad()\n",
        "            outputs = model(images)\n",
        "            pred = torch.argmax(outputs, 1).to('cpu')\n",
        "            loss = criterion(outputs, labels)\n",
        "\n",
        "            test_acc += (pred == cpu_labels).sum().item()\n",
        "            test_loss += loss\n",
        "            test_f1 = f1_score(pred, cpu_labels, average='macro')\n",
        "    test_acc /= len(test_set)\n",
        "    print(f'Epoch{epoch+1}\\t\\tLoss\\tAcc\\tF1')\n",
        "    print(f\"Train:\\t\\t{'%.4f'%train_loss}\\t{'%.4f'%train_acc}\\t{'%.4f'%train_f1}\")\n",
        "    print(f\"Test:\\t\\t{'%.4f'%test_loss}\\t{'%.4f'%test_acc}\\t{'%.4f'%test_f1}\")\n",
        "    print()\n",
        "    metrics['train loss'].append(train_loss.item())\n",
        "    metrics['train acc'].append(train_acc)\n",
        "    metrics['train f1'].append(train_f1)\n",
        "    metrics['test loss'].append(test_loss.item())\n",
        "    metrics['test acc'].append(test_acc)\n",
        "    metrics['test f1'].append(test_f1)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "9IDCxXmiKZH1",
        "outputId": "69a3d83f-a62b-4c8b-ad86-22395265f18f"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch1\t\tLoss\tAcc\tF1\n",
            "Train:\t\t2.9066\t0.1010\t0.0975\n",
            "Test:\t\t2.3586\t0.2000\t0.1271\n",
            "\n",
            "Epoch2\t\tLoss\tAcc\tF1\n",
            "Train:\t\t1.7754\t0.3665\t0.3602\n",
            "Test:\t\t1.2957\t0.5491\t0.5289\n",
            "\n",
            "Epoch3\t\tLoss\tAcc\tF1\n",
            "Train:\t\t1.1232\t0.6106\t0.6076\n",
            "Test:\t\t0.9903\t0.6533\t0.6415\n",
            "\n",
            "Epoch4\t\tLoss\tAcc\tF1\n",
            "Train:\t\t0.8326\t0.7160\t0.7148\n",
            "Test:\t\t0.9442\t0.6860\t0.6730\n",
            "\n",
            "Epoch5\t\tLoss\tAcc\tF1\n",
            "Train:\t\t0.6817\t0.7717\t0.7710\n",
            "Test:\t\t0.6638\t0.7708\t0.7695\n",
            "\n",
            "Epoch6\t\tLoss\tAcc\tF1\n",
            "Train:\t\t0.5840\t0.8057\t0.8052\n",
            "Test:\t\t0.5224\t0.8292\t0.8281\n",
            "\n",
            "Epoch7\t\tLoss\tAcc\tF1\n",
            "Train:\t\t0.5290\t0.8234\t0.8231\n",
            "Test:\t\t0.5228\t0.8256\t0.8225\n",
            "\n",
            "Epoch8\t\tLoss\tAcc\tF1\n",
            "Train:\t\t0.4726\t0.8399\t0.8398\n",
            "Test:\t\t0.4760\t0.8381\t0.8374\n",
            "\n",
            "Epoch9\t\tLoss\tAcc\tF1\n",
            "Train:\t\t0.4344\t0.8564\t0.8563\n",
            "Test:\t\t0.4731\t0.8405\t0.8385\n",
            "\n",
            "Epoch10\t\tLoss\tAcc\tF1\n",
            "Train:\t\t0.4025\t0.8645\t0.8645\n",
            "Test:\t\t0.4369\t0.8545\t0.8551\n",
            "\n",
            "Epoch11\t\tLoss\tAcc\tF1\n",
            "Train:\t\t0.3824\t0.8736\t0.8735\n",
            "Test:\t\t0.3804\t0.8714\t0.8713\n",
            "\n",
            "Epoch12\t\tLoss\tAcc\tF1\n",
            "Train:\t\t0.3573\t0.8769\t0.8768\n",
            "Test:\t\t0.4139\t0.8613\t0.8620\n",
            "\n",
            "Epoch13\t\tLoss\tAcc\tF1\n",
            "Train:\t\t0.3389\t0.8862\t0.8862\n",
            "Test:\t\t0.3660\t0.8812\t0.8801\n",
            "\n",
            "Epoch14\t\tLoss\tAcc\tF1\n",
            "Train:\t\t0.3189\t0.8923\t0.8923\n",
            "Test:\t\t0.3727\t0.8792\t0.8778\n",
            "\n",
            "Epoch15\t\tLoss\tAcc\tF1\n",
            "Train:\t\t0.3072\t0.8956\t0.8956\n",
            "Test:\t\t0.3540\t0.8851\t0.8853\n",
            "\n",
            "Epoch16\t\tLoss\tAcc\tF1\n",
            "Train:\t\t0.2955\t0.9002\t0.9002\n",
            "Test:\t\t0.3634\t0.8792\t0.8800\n",
            "\n",
            "Epoch17\t\tLoss\tAcc\tF1\n",
            "Train:\t\t0.2864\t0.9020\t0.9020\n",
            "Test:\t\t0.3400\t0.8943\t0.8938\n",
            "\n",
            "Epoch18\t\tLoss\tAcc\tF1\n",
            "Train:\t\t0.2699\t0.9086\t0.9086\n",
            "Test:\t\t0.3318\t0.8896\t0.8897\n",
            "\n",
            "Epoch19\t\tLoss\tAcc\tF1\n",
            "Train:\t\t0.2685\t0.9073\t0.9072\n",
            "Test:\t\t0.3041\t0.9000\t0.8997\n",
            "\n",
            "Epoch20\t\tLoss\tAcc\tF1\n",
            "Train:\t\t0.2494\t0.9132\t0.9132\n",
            "Test:\t\t0.3193\t0.8979\t0.8970\n",
            "\n",
            "Epoch21\t\tLoss\tAcc\tF1\n",
            "Train:\t\t0.2356\t0.9224\t0.9224\n",
            "Test:\t\t0.3123\t0.9027\t0.9033\n",
            "\n",
            "Epoch22\t\tLoss\tAcc\tF1\n",
            "Train:\t\t0.2374\t0.9176\t0.9175\n",
            "Test:\t\t0.3007\t0.9003\t0.8999\n",
            "\n",
            "Epoch23\t\tLoss\tAcc\tF1\n",
            "Train:\t\t0.2217\t0.9236\t0.9235\n",
            "Test:\t\t0.3149\t0.8964\t0.8960\n",
            "\n",
            "Epoch24\t\tLoss\tAcc\tF1\n",
            "Train:\t\t0.2177\t0.9259\t0.9259\n",
            "Test:\t\t0.2959\t0.9065\t0.9068\n",
            "\n",
            "Epoch25\t\tLoss\tAcc\tF1\n",
            "Train:\t\t0.2065\t0.9289\t0.9289\n",
            "Test:\t\t0.3046\t0.9036\t0.9037\n",
            "\n",
            "Epoch26\t\tLoss\tAcc\tF1\n",
            "Train:\t\t0.2020\t0.9286\t0.9286\n",
            "Test:\t\t0.3103\t0.9021\t0.9013\n",
            "\n",
            "Epoch27\t\tLoss\tAcc\tF1\n",
            "Train:\t\t0.1894\t0.9336\t0.9335\n",
            "Test:\t\t0.3041\t0.9018\t0.9021\n",
            "\n",
            "Epoch28\t\tLoss\tAcc\tF1\n",
            "Train:\t\t0.1847\t0.9347\t0.9348\n",
            "Test:\t\t0.3289\t0.8997\t0.8994\n",
            "\n",
            "Epoch29\t\tLoss\tAcc\tF1\n",
            "Train:\t\t0.1822\t0.9367\t0.9367\n",
            "Test:\t\t0.3130\t0.9062\t0.9063\n",
            "\n",
            "Epoch30\t\tLoss\tAcc\tF1\n",
            "Train:\t\t0.1776\t0.9399\t0.9398\n",
            "Test:\t\t0.3471\t0.8914\t0.8907\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Plotting results and confusion matrix"
      ],
      "metadata": {
        "id": "iXxjb-V8OR6u"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "x = [i + 1 for i in range(len( metrics['train loss']))]\n",
        "plt.plot(x, metrics['train loss'], color='r', label='train')\n",
        "plt.plot(x, metrics['test loss'], color='b', label='test')\n",
        "plt.xlabel(\"Epoch\")\n",
        "plt.ylabel(\"Loss\")\n",
        "plt.title(\"Loss vs Epoch\")\n",
        "plt.legend()\n",
        "plt.show()\n",
        "\n",
        "plt.plot(x, metrics['train acc'], color='r', label='train')\n",
        "plt.plot(x, metrics['test acc'], color='b', label='test')\n",
        "plt.xlabel(\"Epoch\")\n",
        "plt.ylabel(\"Accuracy\")\n",
        "plt.title(\"Accuracy vs Epoch\")\n",
        "plt.legend()\n",
        "plt.show()\n",
        "\n",
        "plt.plot(x, metrics['train f1'], color='r', label='train')\n",
        "plt.plot(x, metrics['test f1'], color='b', label='test')\n",
        "plt.xlabel(\"Epoch\")\n",
        "plt.ylabel(\"F1 Score\")\n",
        "plt.title(\"F1 Score vs Epoch\")\n",
        "plt.legend()\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 851
        },
        "id": "vJUPZcc-RKU9",
        "outputId": "e4bdbe8d-7149-44d6-d977-7175ce43f3d9"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "all_preds = []\n",
        "all_labels = []\n",
        "\n",
        "model.eval()\n",
        "with torch.no_grad():\n",
        "    for images, labels in test_loader:\n",
        "        all_labels += labels\n",
        "        images = images.to(device)\n",
        "        optimizer.zero_grad()\n",
        "        outputs = model(images)\n",
        "        preds = torch.argmax(outputs, 1).to('cpu')\n",
        "        all_preds += preds\n",
        "\n",
        "df_cm = pd.DataFrame(confusion_matrix(all_preds,all_labels))\n",
        "plt.figure(figsize = (10,7))\n",
        "sn.heatmap(df_cm, annot=True)\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 432
        },
        "id": "FG7WA6lWbz_I",
        "outputId": "505c5178-375c-48ad-9cd6-1514a156803d"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 720x504 with 2 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Interestingly, a lot of samples from the second class are predicted as 3. This behaviour of the model suggests that classes 2 and 3 are pretty similar, which is indeed true. class2 is \"ب\" and class3 is \"ث\". The same thing has happened for classes 15(\"ط\") and 16(\"ظ\") as well."
      ],
      "metadata": {
        "id": "I399hBEVLK2n"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Conclusion\n",
        "Though we used a very well known model, some Arabic letters are very similar to each other, which can decrease the accuracy of the model."
      ],
      "metadata": {
        "id": "2zs_yFZ_Oe55"
      }
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "Y7WXEtIzcFhb"
      },
      "execution_count": null,
      "outputs": []
    }
  ],
  "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": "3.10.5"
    },
    "colab": {
      "provenance": []
    },
    "accelerator": "GPU",
    "gpuClass": "standard"
  },
  "nbformat": 4,
  "nbformat_minor": 0
}