CIFAR-10 Classification.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Import Dependencies"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Import Dependencies\n",
    "import torch\n",
    "import torchvision\n",
    "import torchvision.transforms as transforms\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Define Image Data Transform"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Define the image transform\n",
    "# The output of torchvision datasets are PILImage images of range [0, 1]. \n",
    "# We transform them to Tensors of normalized range [-1, 1].\n",
    "transform = transforms.Compose([transforms.ToTensor(),\n",
    "                                transforms.Normalize((0.5,0.5,0.5), (0.5,0.5,0.5))])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Load Training and Test Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Files already downloaded and verified\n",
      "Files already downloaded and verified\n"
     ]
    }
   ],
   "source": [
    "# Load Training Data\n",
    "train_data = torchvision.datasets.CIFAR10(root='./cifar_data/', train=True, download=True, transform=transform)\n",
    "\n",
    "# Load Test Data\n",
    "test_data = torchvision.datasets.CIFAR10(root='./cifar_data/', train=False, download=True, transform=transform)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Create Training and Test Data Loader"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Create Train Data Loader\n",
    "trainLoader = torch.utils.data.DataLoader(train_data, batch_size=5, shuffle=True, num_workers=3)\n",
    "\n",
    "# Create Test Data Loader\n",
    "testLoader = torch.utils.data.DataLoader(test_data, batch_size=5, shuffle=False, num_workers=3)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Define Class labels for further use"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Define Class Labels\n",
    "class_labels = ('plane', 'car', 'bird', 'cat', 'deer', 'dog', 'frog', 'horse', 'ship', 'truck')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Plot Sample Images from Dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Function to show sample images\n",
    "def show_image(image=None, label=None):\n",
    "    labels = []\n",
    "    plt.figure(figsize=(20,10))\n",
    "    # Un-normalize the Image\n",
    "    image = image/2 + 0.5\n",
    "    # Convert Image to Numpy for plotting\n",
    "    np_img = image.numpy()\n",
    "    plt.imshow(np.transpose(np_img, (1,2,0)))\n",
    "    \n",
    "    l = label.numpy()\n",
    "    \n",
    "    for i in l:\n",
    "        labels.append(class_labels[i])\n",
    "        \n",
    "    plt.title(\"Labels: \" + str(labels))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Get some random images from training set\n",
    "dataIter = iter(trainLoader)\n",
    "img, labels = dataIter.next()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x720 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Show Images\n",
    "show_image(image=torchvision.utils.make_grid(img), label=labels)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Define a Convolutional Neural Network"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Import Dependencies\n",
    "import torch.nn as nn\n",
    "import torch.nn.functional as F\n",
    "import torch.optim as optim"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "# CNN Model class\n",
    "class CNNModel(nn.Module):\n",
    "    # Initialize all Layers\n",
    "    def __init__(self):\n",
    "        super(CNNModel, self).__init__()\n",
    "        self.conv1 = nn.Conv2d(3, 6, 5)\n",
    "        self.pool = nn.MaxPool2d(2, 2)\n",
    "        self.conv2 = nn.Conv2d(6, 16, 5)\n",
    "        self.fc1 = nn.Linear(16 * 5 * 5, 120)\n",
    "        self.fc2 = nn.Linear(120, 84)\n",
    "        self.fc3 = nn.Linear(84, 10)\n",
    "\n",
    "    def forward(self, x):\n",
    "        x = self.pool(F.relu(self.conv1(x)))\n",
    "        x = self.pool(F.relu(self.conv2(x)))\n",
    "        x = x.view(-1, 16 * 5 * 5)\n",
    "        x = F.relu(self.fc1(x))\n",
    "        x = F.relu(self.fc2(x))\n",
    "        x = self.fc3(x)\n",
    "        return x"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Initialize the Model\n",
    "cnn = CNNModel()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Define Loss Function\n",
    "loss_criterion = nn.CrossEntropyLoss()\n",
    "\n",
    "# Define optimizer\n",
    "optimizer = optim.Adam(cnn.parameters(), lr=0.001)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Train the Network"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch: 1\t i: 2000\t Loss: 1.8343812523782252\n",
      "Epoch: 1\t i: 4000\t Loss: 1.552497156277299\n",
      "Epoch: 1\t i: 6000\t Loss: 1.4457302920371293\n",
      "Epoch: 1\t i: 8000\t Loss: 1.4124322621151806\n",
      "Epoch: 1\t i: 10000\t Loss: 1.3608716068193316\n",
      "Epoch: 2\t i: 2000\t Loss: 1.2771225582435728\n",
      "Epoch: 2\t i: 4000\t Loss: 1.2644084264487028\n",
      "Epoch: 2\t i: 6000\t Loss: 1.2526400725096465\n",
      "Epoch: 2\t i: 8000\t Loss: 1.2414843060672283\n",
      "Epoch: 2\t i: 10000\t Loss: 1.2116613011285662\n",
      "Epoch: 3\t i: 2000\t Loss: 1.1488727328665556\n",
      "Epoch: 3\t i: 4000\t Loss: 1.1544695878177882\n",
      "Epoch: 3\t i: 6000\t Loss: 1.1664792434200644\n",
      "Epoch: 3\t i: 8000\t Loss: 1.1473809243068098\n",
      "Epoch: 3\t i: 10000\t Loss: 1.1287699852846562\n",
      "\n",
      "Finished Training...\n"
     ]
    }
   ],
   "source": [
    "epochs = 3\n",
    "\n",
    "for epoch in range(epochs):\n",
    "    running_loss = 0.0\n",
    "    \n",
    "    for i, data in enumerate(trainLoader, 0):\n",
    "        # Load Images and Labels in Batches\n",
    "        images, labels = data\n",
    "        \n",
    "        # Zero the Optimizer Gradient Values\n",
    "        optimizer.zero_grad()\n",
    "        \n",
    "        # Define Model output\n",
    "        y_hat = cnn(images)\n",
    "        \n",
    "        # Define Loss\n",
    "        loss = loss_criterion(y_hat, labels)\n",
    "        \n",
    "        # Do backward Propagation of gradients\n",
    "        loss.backward()\n",
    "        \n",
    "        # Take one step for optimizer\n",
    "        optimizer.step()\n",
    "        \n",
    "        # Capture Loss\n",
    "        running_loss += loss.item()\n",
    "        \n",
    "        # Print Loss and Accuracy\n",
    "        if i%2000 == 1999:\n",
    "            print(\"Epoch: {0}\\t i: {1}\\t Loss: {2}\".format(epoch+1, i+1, running_loss/2000))\n",
    "            # Reset Running Loss\n",
    "            running_loss = 0.0\n",
    "\n",
    "print(\"\\nFinished Training...\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Plot Sample Test Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Get a batch of data from test set\n",
    "dataIter = iter(testLoader)\n",
    "images, labels = dataIter.next()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x720 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Show Images\n",
    "show_image(image=torchvision.utils.make_grid(images), label=labels)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Make Predictions on the above Images using Trained Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Predicted Labels: ['cat', 'ship', 'plane', 'plane', 'bird']\n"
     ]
    }
   ],
   "source": [
    "y_out = cnn(images)\n",
    "_, y_pred = torch.max(y_out, 1)\n",
    "\n",
    "# Predicted Labels\n",
    "y_pred = y_pred.numpy()\n",
    "pred_labels = []\n",
    "for i in y_pred:\n",
    "    pred_labels.append(class_labels[i])\n",
    "print(\"Predicted Labels: {}\".format(pred_labels))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Make Predictions on Whole Test Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model Accuracy on Test Data: 59.08\n"
     ]
    }
   ],
   "source": [
    "# Num Correct Labels\n",
    "num_correct = 0\n",
    "\n",
    "# Total Predictions\n",
    "total = 0\n",
    "\n",
    "# We don't need to backpropagate or find gradients, so turn it Off\n",
    "with torch.no_grad():\n",
    "    for data in testLoader:\n",
    "        # Get test images and labels in batches\n",
    "        images, labels = data\n",
    "        # Get Model prediction on images\n",
    "        y_out = cnn(images)\n",
    "        # Get max probb value for each prediction made\n",
    "        _, y_pred = torch.max(y_out, 1)\n",
    "        # total number of Labels in test set\n",
    "        total += labels.size(0)\n",
    "        # Number of correct predictions\n",
    "        num_correct += (y_pred == labels).sum().item()\n",
    "print(\"Model Accuracy on Test Data: {}\".format(100*num_correct/total))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Which classes Perfomed Well and Which did not"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy of plane : 61 %\n",
      "Accuracy of   car : 73 %\n",
      "Accuracy of  bird : 54 %\n",
      "Accuracy of   cat : 32 %\n",
      "Accuracy of  deer : 43 %\n",
      "Accuracy of   dog : 45 %\n",
      "Accuracy of  frog : 76 %\n",
      "Accuracy of horse : 64 %\n",
      "Accuracy of  ship : 65 %\n",
      "Accuracy of truck : 72 %\n"
     ]
    }
   ],
   "source": [
    "class_correct = list(0. for i in range(10))\n",
    "class_total = list(0. for i in range(10))\n",
    "\n",
    "with torch.no_grad():\n",
    "    for data in testLoader:\n",
    "        images, labels = data\n",
    "        y_out = cnn(images)\n",
    "        _, y_pred = torch.max(y_out, 1)\n",
    "        c = (y_pred == labels).squeeze()\n",
    "        for i in range(4):\n",
    "            label = labels[i]\n",
    "            class_correct[label] += c[i].item()\n",
    "            class_total[label] += 1\n",
    "\n",
    "\n",
    "for i in range(10):\n",
    "    print('Accuracy of %5s : %2d %%' % (class_labels[i], 100 * class_correct[i] / class_total[i]))"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}