x2_Deeper_FC_Digit_Cassification_mnist.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "device(type='cpu')"
      ]
     },
     "execution_count": 1,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import torch\n",
    "import torch.nn as nn\n",
    "import torch.optim as optim \n",
    "from torch.utils.data import DataLoader, random_split\n",
    "import torchvision.transforms as transforms\n",
    "import torchvision.datasets as datasets\n",
    "\n",
    "import time\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "# Set device\n",
    "device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')\n",
    "device"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "#dataset preparation\n",
    "\n",
    "# Hyperparameters\n",
    "input_size = 784 # 28x28\n",
    "num_classes = 10\n",
    "batch_size = 64\n",
    "\n",
    "# Load Data\n",
    "transform = transforms.Compose([\n",
    "    transforms.ToTensor(),\n",
    "    transforms.Normalize((0.1307,), (0.3081,))   #to first arg is mean value and second arg is standard deviation\n",
    "])\n",
    "\n",
    "# Download and prepare the MNIST dataset\n",
    "train_dataset = datasets.MNIST(root='mnist_dataset/', train=True, transform=transform, download=True)\n",
    "test_dataset = datasets.MNIST(root='mnist_dataset/', train=False, transform=transform, download=True)\n",
    "\n",
    "# Split training dataset into training and validation\n",
    "train_size = int(0.8 * len(train_dataset))\n",
    "val_size = len(train_dataset) - train_size\n",
    "train_dataset, val_dataset = random_split(train_dataset, [train_size, val_size])\n",
    "\n",
    "train_loader = DataLoader(dataset=train_dataset, batch_size=batch_size, shuffle=True)\n",
    "val_loader = DataLoader(dataset=val_dataset, batch_size=batch_size, shuffle=False)\n",
    "test_loader = DataLoader(dataset=test_dataset, batch_size=batch_size, shuffle=False)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Images:\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1500x300 with 5 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Validation Images:\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1500x300 with 5 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Function to visualize 5 random images with their labels\n",
    "def visualize_random_images(dataset):\n",
    "    fig, axes = plt.subplots(1, 5, figsize=(15, 3))\n",
    "    for ax in axes:\n",
    "        index = np.random.randint(0, len(dataset))\n",
    "        img, label = dataset[index]\n",
    "        img = img.numpy().squeeze()\n",
    "        ax.imshow(img, cmap='gray')\n",
    "        ax.set_title(f'Label: {label}')\n",
    "        ax.axis('off')\n",
    "    plt.show()\n",
    "\n",
    "# Visualize training and validation images\n",
    "print(\"Training Images:\")\n",
    "visualize_random_images(train_dataset)\n",
    "print(\"Validation Images:\")\n",
    "visualize_random_images(val_dataset)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Function for training the model\n",
    "def train(model, loader, criterion, optimizer):\n",
    "    model.train()\n",
    "    running_loss = 0\n",
    "    correct = 0\n",
    "    total = 0\n",
    "    for data, targets in loader:\n",
    "        data = data.to(device)\n",
    "        targets = targets.to(device)\n",
    "        \n",
    "        # Forward pass\n",
    "        scores = model(data)\n",
    "        loss = criterion(scores, targets)\n",
    "        \n",
    "        # Backward pass\n",
    "        optimizer.zero_grad()\n",
    "        loss.backward()\n",
    "        optimizer.step()\n",
    "        \n",
    "        # Statistics\n",
    "        running_loss += loss.item()\n",
    "        _, predictions = scores.max(1)\n",
    "        correct += (predictions == targets).sum().item()\n",
    "        total += targets.size(0)\n",
    "    \n",
    "    epoch_loss = running_loss / len(loader)\n",
    "    epoch_acc = correct / total\n",
    "    return epoch_loss, epoch_acc\n",
    "\n",
    "# Function for evaluating the model\n",
    "def evaluate(model, loader, criterion):\n",
    "    model.eval()\n",
    "    running_loss = 0\n",
    "    correct = 0\n",
    "    total = 0\n",
    "    with torch.no_grad():\n",
    "        for data, targets in loader:\n",
    "            data = data.to(device)\n",
    "            targets = targets.to(device)\n",
    "            \n",
    "            # Forward pass\n",
    "            scores = model(data)\n",
    "            loss = criterion(scores, targets)\n",
    "            \n",
    "            # Statistics\n",
    "            running_loss += loss.item()\n",
    "            _, predictions = scores.max(1)\n",
    "            correct += (predictions == targets).sum().item()\n",
    "            total += targets.size(0)\n",
    "    \n",
    "    epoch_loss = running_loss / len(loader)\n",
    "    epoch_acc = correct / total\n",
    "    return epoch_loss, epoch_acc"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Define the Neural Network\n",
    "class deeper_fc(nn.Module):\n",
    "    def __init__(self, input_size, num_classes):\n",
    "        super(deeper_fc, self).__init__()\n",
    "        self.fc1 = nn.Linear(input_size, 256)\n",
    "        self.fc2 = nn.Linear(256, 128)\n",
    "        self.fc3 = nn.Linear(128, 64)\n",
    "        self.fc4 = nn.Linear(64, 32)\n",
    "        self.fc5 = nn.Linear(32, num_classes)\n",
    "    \n",
    "    def forward(self, x):\n",
    "        x = x.reshape(x.shape[0], -1)\n",
    "        x = torch.relu(self.fc1(x))\n",
    "        x = torch.relu(self.fc2(x))\n",
    "        x = torch.relu(self.fc3(x))\n",
    "        x = torch.relu(self.fc4(x))\n",
    "        x = self.fc5(x)\n",
    "        return x\n",
    "\n",
    "# Initialize network\n",
    "model = deeper_fc(input_size=input_size, num_classes=num_classes).to(device)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "#hyperparameter\n",
    "learning_rate = 0.001\n",
    "\n",
    "# Loss and optimizer\n",
    "criterion = nn.CrossEntropyLoss()\n",
    "optimizer = optim.Adam(model.parameters(), lr=learning_rate)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Training loop\n",
    "train_losses = []\n",
    "val_losses = []\n",
    "train_accuracies = []\n",
    "val_accuracies = []\n",
    "\n",
    "num_epochs = 100\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch [1/100]: Train Loss: 0.3399, Train Acc: 0.8950, Val Loss: 0.1442, Val Acc: 0.9555\n",
      "Epoch [2/100]: Train Loss: 0.1278, Train Acc: 0.9609, Val Loss: 0.1177, Val Acc: 0.9640\n",
      "Epoch [3/100]: Train Loss: 0.0885, Train Acc: 0.9729, Val Loss: 0.0946, Val Acc: 0.9702\n",
      "Epoch [4/100]: Train Loss: 0.0660, Train Acc: 0.9795, Val Loss: 0.1034, Val Acc: 0.9690\n",
      "Epoch [5/100]: Train Loss: 0.0529, Train Acc: 0.9827, Val Loss: 0.0874, Val Acc: 0.9747\n",
      "Epoch [6/100]: Train Loss: 0.0427, Train Acc: 0.9866, Val Loss: 0.0946, Val Acc: 0.9733\n",
      "Epoch [7/100]: Train Loss: 0.0380, Train Acc: 0.9876, Val Loss: 0.1059, Val Acc: 0.9712\n",
      "Epoch [8/100]: Train Loss: 0.0336, Train Acc: 0.9891, Val Loss: 0.0815, Val Acc: 0.9782\n",
      "Epoch [9/100]: Train Loss: 0.0289, Train Acc: 0.9906, Val Loss: 0.1119, Val Acc: 0.9724\n",
      "Epoch [10/100]: Train Loss: 0.0255, Train Acc: 0.9919, Val Loss: 0.0999, Val Acc: 0.9747\n",
      "Epoch [11/100]: Train Loss: 0.0240, Train Acc: 0.9917, Val Loss: 0.0898, Val Acc: 0.9778\n",
      "Epoch [12/100]: Train Loss: 0.0251, Train Acc: 0.9924, Val Loss: 0.0954, Val Acc: 0.9775\n",
      "Epoch [13/100]: Train Loss: 0.0217, Train Acc: 0.9931, Val Loss: 0.1073, Val Acc: 0.9766\n",
      "Epoch [14/100]: Train Loss: 0.0173, Train Acc: 0.9945, Val Loss: 0.0861, Val Acc: 0.9793\n",
      "Epoch [15/100]: Train Loss: 0.0164, Train Acc: 0.9950, Val Loss: 0.0896, Val Acc: 0.9806\n",
      "Epoch [16/100]: Train Loss: 0.0176, Train Acc: 0.9944, Val Loss: 0.1119, Val Acc: 0.9759\n",
      "Epoch [17/100]: Train Loss: 0.0178, Train Acc: 0.9945, Val Loss: 0.0901, Val Acc: 0.9791\n",
      "Epoch [18/100]: Train Loss: 0.0166, Train Acc: 0.9946, Val Loss: 0.1062, Val Acc: 0.9759\n",
      "Epoch [19/100]: Train Loss: 0.0136, Train Acc: 0.9958, Val Loss: 0.1195, Val Acc: 0.9772\n",
      "Epoch [20/100]: Train Loss: 0.0131, Train Acc: 0.9958, Val Loss: 0.1276, Val Acc: 0.9758\n",
      "Epoch [21/100]: Train Loss: 0.0138, Train Acc: 0.9959, Val Loss: 0.1166, Val Acc: 0.9772\n",
      "Epoch [22/100]: Train Loss: 0.0125, Train Acc: 0.9966, Val Loss: 0.1206, Val Acc: 0.9763\n",
      "Epoch [23/100]: Train Loss: 0.0124, Train Acc: 0.9960, Val Loss: 0.1129, Val Acc: 0.9778\n",
      "Epoch [24/100]: Train Loss: 0.0135, Train Acc: 0.9964, Val Loss: 0.0962, Val Acc: 0.9815\n",
      "Epoch [25/100]: Train Loss: 0.0109, Train Acc: 0.9968, Val Loss: 0.1163, Val Acc: 0.9766\n",
      "Epoch [26/100]: Train Loss: 0.0103, Train Acc: 0.9970, Val Loss: 0.1026, Val Acc: 0.9806\n",
      "Epoch [27/100]: Train Loss: 0.0110, Train Acc: 0.9971, Val Loss: 0.1132, Val Acc: 0.9782\n",
      "Epoch [28/100]: Train Loss: 0.0104, Train Acc: 0.9969, Val Loss: 0.1143, Val Acc: 0.9799\n",
      "Epoch [29/100]: Train Loss: 0.0119, Train Acc: 0.9963, Val Loss: 0.1160, Val Acc: 0.9778\n",
      "Epoch [30/100]: Train Loss: 0.0096, Train Acc: 0.9974, Val Loss: 0.1194, Val Acc: 0.9775\n",
      "Epoch [31/100]: Train Loss: 0.0108, Train Acc: 0.9972, Val Loss: 0.1049, Val Acc: 0.9790\n",
      "Epoch [32/100]: Train Loss: 0.0092, Train Acc: 0.9972, Val Loss: 0.0975, Val Acc: 0.9802\n",
      "Epoch [33/100]: Train Loss: 0.0063, Train Acc: 0.9981, Val Loss: 0.1238, Val Acc: 0.9784\n",
      "Epoch [34/100]: Train Loss: 0.0096, Train Acc: 0.9973, Val Loss: 0.1018, Val Acc: 0.9817\n",
      "Epoch [35/100]: Train Loss: 0.0096, Train Acc: 0.9971, Val Loss: 0.1263, Val Acc: 0.9794\n",
      "Epoch [36/100]: Train Loss: 0.0054, Train Acc: 0.9984, Val Loss: 0.1232, Val Acc: 0.9789\n",
      "Epoch [37/100]: Train Loss: 0.0084, Train Acc: 0.9974, Val Loss: 0.1320, Val Acc: 0.9798\n",
      "Epoch [38/100]: Train Loss: 0.0091, Train Acc: 0.9977, Val Loss: 0.1179, Val Acc: 0.9808\n",
      "Epoch [39/100]: Train Loss: 0.0044, Train Acc: 0.9987, Val Loss: 0.1170, Val Acc: 0.9804\n",
      "Epoch [40/100]: Train Loss: 0.0084, Train Acc: 0.9976, Val Loss: 0.1209, Val Acc: 0.9788\n",
      "Epoch [41/100]: Train Loss: 0.0073, Train Acc: 0.9978, Val Loss: 0.1299, Val Acc: 0.9792\n",
      "Epoch [42/100]: Train Loss: 0.0092, Train Acc: 0.9974, Val Loss: 0.1120, Val Acc: 0.9808\n",
      "Epoch [43/100]: Train Loss: 0.0063, Train Acc: 0.9982, Val Loss: 0.1425, Val Acc: 0.9777\n",
      "Epoch [44/100]: Train Loss: 0.0074, Train Acc: 0.9980, Val Loss: 0.2033, Val Acc: 0.9717\n",
      "Epoch [45/100]: Train Loss: 0.0079, Train Acc: 0.9979, Val Loss: 0.1300, Val Acc: 0.9802\n",
      "Epoch [46/100]: Train Loss: 0.0074, Train Acc: 0.9976, Val Loss: 0.1356, Val Acc: 0.9772\n",
      "Epoch [47/100]: Train Loss: 0.0044, Train Acc: 0.9986, Val Loss: 0.1603, Val Acc: 0.9798\n",
      "Epoch [48/100]: Train Loss: 0.0070, Train Acc: 0.9980, Val Loss: 0.1291, Val Acc: 0.9796\n",
      "Epoch [49/100]: Train Loss: 0.0066, Train Acc: 0.9980, Val Loss: 0.1125, Val Acc: 0.9825\n",
      "Epoch [50/100]: Train Loss: 0.0057, Train Acc: 0.9986, Val Loss: 0.1065, Val Acc: 0.9828\n",
      "Epoch [51/100]: Train Loss: 0.0077, Train Acc: 0.9978, Val Loss: 0.1309, Val Acc: 0.9802\n",
      "Epoch [52/100]: Train Loss: 0.0050, Train Acc: 0.9983, Val Loss: 0.1253, Val Acc: 0.9829\n",
      "Epoch [53/100]: Train Loss: 0.0067, Train Acc: 0.9985, Val Loss: 0.1315, Val Acc: 0.9802\n",
      "Epoch [54/100]: Train Loss: 0.0060, Train Acc: 0.9982, Val Loss: 0.1490, Val Acc: 0.9776\n",
      "Epoch [55/100]: Train Loss: 0.0089, Train Acc: 0.9979, Val Loss: 0.1256, Val Acc: 0.9790\n",
      "Epoch [56/100]: Train Loss: 0.0063, Train Acc: 0.9984, Val Loss: 0.1184, Val Acc: 0.9815\n",
      "Epoch [57/100]: Train Loss: 0.0032, Train Acc: 0.9991, Val Loss: 0.1165, Val Acc: 0.9822\n",
      "Epoch [58/100]: Train Loss: 0.0061, Train Acc: 0.9985, Val Loss: 0.1277, Val Acc: 0.9809\n",
      "Epoch [59/100]: Train Loss: 0.0071, Train Acc: 0.9982, Val Loss: 0.1226, Val Acc: 0.9797\n",
      "Epoch [60/100]: Train Loss: 0.0058, Train Acc: 0.9982, Val Loss: 0.1403, Val Acc: 0.9811\n",
      "Epoch [61/100]: Train Loss: 0.0047, Train Acc: 0.9988, Val Loss: 0.1469, Val Acc: 0.9801\n",
      "Epoch [62/100]: Train Loss: 0.0079, Train Acc: 0.9979, Val Loss: 0.1405, Val Acc: 0.9815\n",
      "Epoch [63/100]: Train Loss: 0.0034, Train Acc: 0.9991, Val Loss: 0.1559, Val Acc: 0.9810\n",
      "Epoch [64/100]: Train Loss: 0.0086, Train Acc: 0.9979, Val Loss: 0.1343, Val Acc: 0.9799\n",
      "Epoch [65/100]: Train Loss: 0.0057, Train Acc: 0.9982, Val Loss: 0.1431, Val Acc: 0.9812\n",
      "Epoch [66/100]: Train Loss: 0.0046, Train Acc: 0.9988, Val Loss: 0.1187, Val Acc: 0.9826\n",
      "Epoch [67/100]: Train Loss: 0.0008, Train Acc: 0.9999, Val Loss: 0.1260, Val Acc: 0.9833\n",
      "Epoch [68/100]: Train Loss: 0.0084, Train Acc: 0.9979, Val Loss: 0.1406, Val Acc: 0.9783\n",
      "Epoch [69/100]: Train Loss: 0.0047, Train Acc: 0.9988, Val Loss: 0.1245, Val Acc: 0.9821\n",
      "Epoch [70/100]: Train Loss: 0.0059, Train Acc: 0.9984, Val Loss: 0.1418, Val Acc: 0.9792\n",
      "Epoch [71/100]: Train Loss: 0.0045, Train Acc: 0.9987, Val Loss: 0.1663, Val Acc: 0.9801\n",
      "Epoch [72/100]: Train Loss: 0.0060, Train Acc: 0.9982, Val Loss: 0.1317, Val Acc: 0.9825\n",
      "Epoch [73/100]: Train Loss: 0.0031, Train Acc: 0.9993, Val Loss: 0.1725, Val Acc: 0.9769\n",
      "Epoch [74/100]: Train Loss: 0.0093, Train Acc: 0.9979, Val Loss: 0.1439, Val Acc: 0.9809\n",
      "Epoch [75/100]: Train Loss: 0.0035, Train Acc: 0.9991, Val Loss: 0.1521, Val Acc: 0.9822\n",
      "Epoch [76/100]: Train Loss: 0.0024, Train Acc: 0.9995, Val Loss: 0.1823, Val Acc: 0.9789\n",
      "Epoch [77/100]: Train Loss: 0.0087, Train Acc: 0.9979, Val Loss: 0.1488, Val Acc: 0.9792\n",
      "Epoch [78/100]: Train Loss: 0.0053, Train Acc: 0.9987, Val Loss: 0.1384, Val Acc: 0.9810\n",
      "Epoch [79/100]: Train Loss: 0.0033, Train Acc: 0.9992, Val Loss: 0.1512, Val Acc: 0.9814\n",
      "Epoch [80/100]: Train Loss: 0.0051, Train Acc: 0.9989, Val Loss: 0.1247, Val Acc: 0.9815\n",
      "Epoch [81/100]: Train Loss: 0.0047, Train Acc: 0.9989, Val Loss: 0.1600, Val Acc: 0.9796\n",
      "Epoch [82/100]: Train Loss: 0.0076, Train Acc: 0.9982, Val Loss: 0.1580, Val Acc: 0.9804\n",
      "Epoch [83/100]: Train Loss: 0.0022, Train Acc: 0.9994, Val Loss: 0.1583, Val Acc: 0.9825\n",
      "Epoch [84/100]: Train Loss: 0.0036, Train Acc: 0.9991, Val Loss: 0.1644, Val Acc: 0.9801\n",
      "Epoch [85/100]: Train Loss: 0.0055, Train Acc: 0.9986, Val Loss: 0.1563, Val Acc: 0.9798\n",
      "Epoch [86/100]: Train Loss: 0.0072, Train Acc: 0.9982, Val Loss: 0.1491, Val Acc: 0.9799\n",
      "Epoch [87/100]: Train Loss: 0.0021, Train Acc: 0.9994, Val Loss: 0.1444, Val Acc: 0.9825\n",
      "Epoch [88/100]: Train Loss: 0.0020, Train Acc: 0.9996, Val Loss: 0.1641, Val Acc: 0.9812\n",
      "Epoch [89/100]: Train Loss: 0.0081, Train Acc: 0.9982, Val Loss: 0.1562, Val Acc: 0.9795\n",
      "Epoch [90/100]: Train Loss: 0.0048, Train Acc: 0.9987, Val Loss: 0.1474, Val Acc: 0.9818\n",
      "Epoch [91/100]: Train Loss: 0.0040, Train Acc: 0.9990, Val Loss: 0.1611, Val Acc: 0.9811\n",
      "Epoch [92/100]: Train Loss: 0.0049, Train Acc: 0.9988, Val Loss: 0.1544, Val Acc: 0.9803\n",
      "Epoch [93/100]: Train Loss: 0.0042, Train Acc: 0.9988, Val Loss: 0.1582, Val Acc: 0.9794\n",
      "Epoch [94/100]: Train Loss: 0.0031, Train Acc: 0.9991, Val Loss: 0.1778, Val Acc: 0.9813\n",
      "Epoch [95/100]: Train Loss: 0.0002, Train Acc: 1.0000, Val Loss: 0.1590, Val Acc: 0.9826\n",
      "Epoch [96/100]: Train Loss: 0.0000, Train Acc: 1.0000, Val Loss: 0.1649, Val Acc: 0.9828\n",
      "Epoch [97/100]: Train Loss: 0.0000, Train Acc: 1.0000, Val Loss: 0.1683, Val Acc: 0.9830\n",
      "Epoch [98/100]: Train Loss: 0.0000, Train Acc: 1.0000, Val Loss: 0.1732, Val Acc: 0.9828\n",
      "Epoch [99/100]: Train Loss: 0.0000, Train Acc: 1.0000, Val Loss: 0.1777, Val Acc: 0.9827\n",
      "Epoch [100/100]: Train Loss: 0.0000, Train Acc: 1.0000, Val Loss: 0.1823, Val Acc: 0.9826\n",
      "Total training time: 389.84 seconds\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1200x400 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Record the start time\n",
    "start_time = time.time()\n",
    "\n",
    "#start training\n",
    "for epoch in range(num_epochs):\n",
    "    train_loss, train_acc = train(model, train_loader, criterion, optimizer)\n",
    "    val_loss, val_acc = evaluate(model, val_loader, criterion)\n",
    "    \n",
    "    train_losses.append(train_loss)\n",
    "    val_losses.append(val_loss)\n",
    "    train_accuracies.append(train_acc)\n",
    "    val_accuracies.append(val_acc)\n",
    "    \n",
    "    print(f'Epoch [{epoch+1}/{num_epochs}]: Train Loss: {train_loss:.4f}, Train Acc: {train_acc:.4f}, Val Loss: {val_loss:.4f}, Val Acc: {val_acc:.4f}')\n",
    "\n",
    "# Record the end time\n",
    "end_time = time.time()\n",
    "total_time = end_time - start_time\n",
    "\n",
    "print(f'Total training time: {total_time:.2f} seconds')\n",
    "\n",
    "# Plotting the training and validation loss and accuracy\n",
    "epochs = range(1, num_epochs + 1)\n",
    "\n",
    "plt.figure(figsize=(12, 4))\n",
    "\n",
    "plt.subplot(1, 2, 1)\n",
    "plt.plot(epochs, train_losses, label='Train Loss')\n",
    "plt.plot(epochs, val_losses, label='Val Loss')\n",
    "plt.xlabel('Epochs')\n",
    "plt.ylabel('Loss')\n",
    "plt.legend()\n",
    "plt.title('Loss over Epochs')\n",
    "\n",
    "plt.subplot(1, 2, 2)\n",
    "plt.plot(epochs, train_accuracies, label='Train Accuracy')\n",
    "plt.plot(epochs, val_accuracies, label='Val Accuracy')\n",
    "plt.xlabel('Epochs')\n",
    "plt.ylabel('Accuracy')\n",
    "plt.legend()\n",
    "plt.title('Accuracy over Epochs')\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Test Loss: 0.1831, Test Accuracy: 0.9829\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1500x300 with 5 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1500x300 with 5 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Evaluate the model on test data and collect predictions\n",
    "def evaluate_and_collect_predictions(model, loader, criterion):\n",
    "    model.eval()\n",
    "    running_loss = 0\n",
    "    correct = 0\n",
    "    total = 0\n",
    "    correct_images = []\n",
    "    incorrect_images = []\n",
    "    correct_labels = []\n",
    "    incorrect_labels = []\n",
    "    predicted_labels_correct = []\n",
    "    predicted_labels_incorrect = []\n",
    "\n",
    "    with torch.no_grad():\n",
    "        for data, targets in loader:\n",
    "            data = data.to(device)\n",
    "            targets = targets.to(device)\n",
    "            \n",
    "            # Forward pass\n",
    "            outputs = model(data)\n",
    "            loss = criterion(outputs, targets)\n",
    "            \n",
    "            # Statistics\n",
    "            running_loss += loss.item()\n",
    "            _, predicted = outputs.max(1)\n",
    "            correct += (predicted == targets).sum().item()\n",
    "            total += targets.size(0)\n",
    "\n",
    "            for i in range(data.size(0)):\n",
    "                if predicted[i] == targets[i]:\n",
    "                    correct_images.append(data[i].cpu())\n",
    "                    correct_labels.append(targets[i].cpu())\n",
    "                    predicted_labels_correct.append(predicted[i].cpu())\n",
    "                else:\n",
    "                    incorrect_images.append(data[i].cpu())\n",
    "                    incorrect_labels.append(targets[i].cpu())\n",
    "                    predicted_labels_incorrect.append(predicted[i].cpu())\n",
    "\n",
    "    epoch_loss = running_loss / len(loader)\n",
    "    epoch_acc = correct / total\n",
    "\n",
    "    return epoch_loss, epoch_acc, correct_images, incorrect_images, correct_labels, incorrect_labels, predicted_labels_correct, predicted_labels_incorrect\n",
    "\n",
    "# Call the modified evaluation function\n",
    "test_loss, test_acc, correct_images, incorrect_images, correct_labels, incorrect_labels, predicted_labels_correct, predicted_labels_incorrect = evaluate_and_collect_predictions(model, test_loader, criterion)\n",
    "print(f'Test Loss: {test_loss:.4f}, Test Accuracy: {test_acc:.4f}')\n",
    "\n",
    "# Function to visualize 5 random correct and incorrect predictions\n",
    "def visualize_random_predictions(correct_images, incorrect_images, correct_labels, incorrect_labels, predicted_labels_correct, predicted_labels_incorrect):\n",
    "    fig, axes = plt.subplots(1, 5, figsize=(15, 3))\n",
    "    for ax in axes:\n",
    "        idx = np.random.randint(0, len(correct_images))\n",
    "        img = correct_images[idx].numpy().squeeze()\n",
    "        true_label = correct_labels[idx].item()\n",
    "        pred_label = predicted_labels_correct[idx].item()\n",
    "        ax.imshow(img, cmap='gray')\n",
    "        ax.set_title(f'True: {true_label}, Pred: {pred_label}')\n",
    "        ax.axis('off')\n",
    "    plt.suptitle('Correct Predictions')\n",
    "    plt.show()\n",
    "\n",
    "    fig, axes = plt.subplots(1, 5, figsize=(15, 3))\n",
    "    for ax in axes:\n",
    "        idx = np.random.randint(0, len(incorrect_images))\n",
    "        img = incorrect_images[idx].numpy().squeeze()\n",
    "        true_label = incorrect_labels[idx].item()\n",
    "        pred_label = predicted_labels_incorrect[idx].item()\n",
    "        ax.imshow(img, cmap='gray')\n",
    "        ax.set_title(f'True: {true_label}, Pred: {pred_label}')\n",
    "        ax.axis('off')\n",
    "    plt.suptitle('Incorrect Predictions')\n",
    "    plt.show()\n",
    "\n",
    "# Visualize the predictions\n",
    "visualize_random_predictions(correct_images, incorrect_images, correct_labels, incorrect_labels, predicted_labels_correct, predicted_labels_incorrect)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "dl",
   "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": "3.11.8"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}