5.Pytorch FashionMNIST Image Classification using CNN.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "source": [
    "### Fashion MNIST using CNN with Pytorch\r\n",
    "\r\n",
    "### 1.Load Libraries"
   ],
   "metadata": {}
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "source": [
    "# torch loaded...!!!\r\n",
    "import torch\r\n",
    "import torch.nn.functional as F\r\n",
    "import torch.nn as nn\r\n",
    "import torch.optim as optim\r\n",
    "\r\n",
    "# torchvision loaded...!!!\r\n",
    "import torchvision\r\n",
    "import torchvision.transforms as transforms\r\n",
    "from torchsummary import summary\r\n",
    "\r\n",
    "# other libraries loaded...!!!\r\n",
    "import time\r\n",
    "import numpy as np\r\n",
    "import matplotlib.pyplot as plt"
   ],
   "outputs": [],
   "metadata": {}
  },
  {
   "cell_type": "markdown",
   "source": [
    "### 2.Defining some constant"
   ],
   "metadata": {}
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "source": [
    "# Constant defined\r\n",
    "NUM_EPOCHS = 10\r\n",
    "BATCH_SIZE = 4\r\n",
    "LEARNING_RATE = 0.001"
   ],
   "outputs": [],
   "metadata": {}
  },
  {
   "cell_type": "markdown",
   "source": [
    "### 3.Defining the transform"
   ],
   "metadata": {}
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "source": [
    "transform = transforms.Compose([\r\n",
    "    transforms.ToTensor(),\r\n",
    "    transforms.Normalize((0.5,),(0.5,))\r\n",
    "    ])"
   ],
   "outputs": [],
   "metadata": {}
  },
  {
   "cell_type": "markdown",
   "source": [
    "### 4.Download the Data and Load in to DataLoader"
   ],
   "metadata": {}
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "source": [
    "# Fashion-mnist dataset dataset : https://github.com/zalandoresearch/fashion-mnist\r\n",
    "trainset = torchvision.datasets.FashionMNIST(root='./data_new', train=True, download=True, transform=transform)\r\n",
    "testset = torchvision.datasets.FashionMNIST(root='./data_new', train=False, download=True, transform=transform)\r\n",
    "\r\n",
    "# Dataloader\r\n",
    "trainloader = torch.utils.data.DataLoader(trainset, batch_size=BATCH_SIZE, shuffle=True, num_workers=2)\r\n",
    "testloader = torch.utils.data.DataLoader(testset, batch_size=BATCH_SIZE, shuffle=False, num_workers=2)"
   ],
   "outputs": [
    {
     "output_type": "stream",
     "name": "stdout",
     "text": [
      "Downloading http://fashion-mnist.s3-website.eu-central-1.amazonaws.com/train-images-idx3-ubyte.gz\n",
      "Downloading http://fashion-mnist.s3-website.eu-central-1.amazonaws.com/train-images-idx3-ubyte.gz to ./data_new\\FashionMNIST\\raw\\train-images-idx3-ubyte.gz\n"
     ]
    },
    {
     "output_type": "stream",
     "name": "stderr",
     "text": [
      "26422272it [00:07, 3572238.27it/s]                              \n"
     ]
    },
    {
     "output_type": "stream",
     "name": "stdout",
     "text": [
      "Extracting ./data_new\\FashionMNIST\\raw\\train-images-idx3-ubyte.gz to ./data_new\\FashionMNIST\\raw\n",
      "\n",
      "Downloading http://fashion-mnist.s3-website.eu-central-1.amazonaws.com/train-labels-idx1-ubyte.gz\n",
      "Downloading http://fashion-mnist.s3-website.eu-central-1.amazonaws.com/train-labels-idx1-ubyte.gz to ./data_new\\FashionMNIST\\raw\\train-labels-idx1-ubyte.gz\n"
     ]
    },
    {
     "output_type": "stream",
     "name": "stderr",
     "text": [
      "29696it [00:00, 214656.21it/s]                          \n"
     ]
    },
    {
     "output_type": "stream",
     "name": "stdout",
     "text": [
      "Extracting ./data_new\\FashionMNIST\\raw\\train-labels-idx1-ubyte.gz to ./data_new\\FashionMNIST\\raw\n",
      "\n",
      "Downloading http://fashion-mnist.s3-website.eu-central-1.amazonaws.com/t10k-images-idx3-ubyte.gz\n",
      "Downloading http://fashion-mnist.s3-website.eu-central-1.amazonaws.com/t10k-images-idx3-ubyte.gz to ./data_new\\FashionMNIST\\raw\\t10k-images-idx3-ubyte.gz\n"
     ]
    },
    {
     "output_type": "stream",
     "name": "stderr",
     "text": [
      "4422656it [00:01, 2733907.83it/s]                             \n"
     ]
    },
    {
     "output_type": "stream",
     "name": "stdout",
     "text": [
      "Extracting ./data_new\\FashionMNIST\\raw\\t10k-images-idx3-ubyte.gz to ./data_new\\FashionMNIST\\raw\n",
      "\n",
      "Downloading http://fashion-mnist.s3-website.eu-central-1.amazonaws.com/t10k-labels-idx1-ubyte.gz\n",
      "Downloading http://fashion-mnist.s3-website.eu-central-1.amazonaws.com/t10k-labels-idx1-ubyte.gz to ./data_new\\FashionMNIST\\raw\\t10k-labels-idx1-ubyte.gz\n"
     ]
    },
    {
     "output_type": "stream",
     "name": "stderr",
     "text": [
      "6144it [00:00, 6166500.07it/s]          "
     ]
    },
    {
     "output_type": "stream",
     "name": "stdout",
     "text": [
      "Extracting ./data_new\\FashionMNIST\\raw\\t10k-labels-idx1-ubyte.gz to ./data_new\\FashionMNIST\\raw\n",
      "\n"
     ]
    },
    {
     "output_type": "stream",
     "name": "stderr",
     "text": [
      "\n",
      "C:\\Users\\SHIVA\\miniconda3\\envs\\pytorch19\\lib\\site-packages\\torchvision\\datasets\\mnist.py:498: UserWarning: The given NumPy array is not writeable, and PyTorch does not support non-writeable tensors. This means you can write to the underlying (supposedly non-writeable) NumPy array using the tensor. You may want to copy the array to protect its data or make it writeable before converting it to a tensor. This type of warning will be suppressed for the rest of this program. (Triggered internally at  ..\\torch\\csrc\\utils\\tensor_numpy.cpp:180.)\n",
      "  return torch.from_numpy(parsed.astype(m[2], copy=False)).view(*s)\n"
     ]
    }
   ],
   "metadata": {}
  },
  {
   "cell_type": "markdown",
   "source": [],
   "metadata": {}
  },
  {
   "cell_type": "markdown",
   "source": [
    "### 5.Visulization the images"
   ],
   "metadata": {}
  },
  {
   "cell_type": "markdown",
   "source": [
    "![](https://i.imgur.com/KdAFXTE.gif)"
   ],
   "metadata": {}
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "source": [
    "classes = ['T-shirt', 'Trouser', 'Pullover','Dress','Coat','Sandal','Shirt','Sneaker','Bag', 'AnkleBoot']"
   ],
   "outputs": [],
   "metadata": {}
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "source": [
    "for batch_1 in trainloader:\r\n",
    "    batch = batch_1\r\n",
    "    break\r\n",
    "\r\n",
    "print(f\"Image shape:{batch[0].shape}\") # Image pixels -> (batch, channel, height, width)\r\n",
    "print(f\"Label shape:{batch[1].shape}\") # Label -> (batch, 1)\r\n",
    "\r\n",
    "plt.figure(figsize=(12,10))\r\n",
    "for i in range(batch[0].shape[0]):\r\n",
    "    plt.subplot(1,4, i+1)\r\n",
    "    plt.title(classes[batch[1][i]])\r\n",
    "    plt.imshow(batch[0][i,0,:,:])\r\n",
    "    plt.axis('off')\r\n",
    "plt.show()"
   ],
   "outputs": [
    {
     "output_type": "stream",
     "name": "stdout",
     "text": [
      "Image shape:torch.Size([4, 1, 28, 28])\n",
      "Label shape:torch.Size([4])\n"
     ]
    },
    {
     "output_type": "display_data",
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 864x720 with 4 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     }
    }
   ],
   "metadata": {}
  },
  {
   "cell_type": "markdown",
   "source": [
    "### 6. Build CNN Architecture"
   ],
   "metadata": {}
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "source": [
    "class LeNet(nn.Module):\r\n",
    "    def __init__(self):\r\n",
    "        super(LeNet, self).__init__()\r\n",
    "        self.conv1 = nn.Conv2d(in_channels=1, out_channels=6, \r\n",
    "                               kernel_size=5)\r\n",
    "        self.conv2 = nn.Conv2d(in_channels=6, out_channels=16, \r\n",
    "                               kernel_size=5)\r\n",
    "        self.fc1 = nn.Linear(in_features=256, out_features=120)\r\n",
    "        self.fc2 = nn.Linear(in_features=120, out_features=84)\r\n",
    "        self.fc3 = nn.Linear(in_features=84, out_features=10)\r\n",
    "        \r\n",
    "    def forward(self, x):\r\n",
    "        x = F.relu(self.conv1(x))\r\n",
    "        x = F.max_pool2d(x, kernel_size=2)\r\n",
    "        x = F.relu(self.conv2(x))\r\n",
    "        x = F.max_pool2d(x, kernel_size=2)\r\n",
    "        x = x.view(x.size(0), -1)\r\n",
    "        x = F.relu(self.fc1(x))\r\n",
    "        x = F.relu(self.fc2(x))\r\n",
    "        x = self.fc3(x)\r\n",
    "        return x\r\n",
    "net = LeNet()\r\n",
    "print(net)"
   ],
   "outputs": [
    {
     "output_type": "stream",
     "name": "stdout",
     "text": [
      "LeNet(\n",
      "  (conv1): Conv2d(1, 6, kernel_size=(5, 5), stride=(1, 1))\n",
      "  (conv2): Conv2d(6, 16, kernel_size=(5, 5), stride=(1, 1))\n",
      "  (fc1): Linear(in_features=256, out_features=120, bias=True)\n",
      "  (fc2): Linear(in_features=120, out_features=84, bias=True)\n",
      "  (fc3): Linear(in_features=84, out_features=10, bias=True)\n",
      ")\n"
     ]
    }
   ],
   "metadata": {}
  },
  {
   "cell_type": "markdown",
   "source": [
    "### 7.Display Model"
   ],
   "metadata": {}
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "source": [
    "device = torch.device(\"cuda:0\" if torch.cuda.is_available() else \"cpu\")\r\n",
    "print(str(summary(net.to(device), (1, 28, 28), depth=1, device=device)))"
   ],
   "outputs": [
    {
     "output_type": "stream",
     "name": "stdout",
     "text": [
      "==========================================================================================\n",
      "Layer (type:depth-idx)                   Output Shape              Param #\n",
      "==========================================================================================\n",
      "├─Conv2d: 1-1                            [-1, 6, 24, 24]           156\n",
      "├─Conv2d: 1-2                            [-1, 16, 8, 8]            2,416\n",
      "├─Linear: 1-3                            [-1, 120]                 30,840\n",
      "├─Linear: 1-4                            [-1, 84]                  10,164\n",
      "├─Linear: 1-5                            [-1, 10]                  850\n",
      "==========================================================================================\n",
      "Total params: 44,426\n",
      "Trainable params: 44,426\n",
      "Non-trainable params: 0\n",
      "Total mult-adds (M): 0.28\n",
      "==========================================================================================\n",
      "Input size (MB): 0.00\n",
      "Forward/backward pass size (MB): 0.04\n",
      "Params size (MB): 0.17\n",
      "Estimated Total Size (MB): 0.21\n",
      "==========================================================================================\n",
      "==========================================================================================\n",
      "Layer (type:depth-idx)                   Output Shape              Param #\n",
      "==========================================================================================\n",
      "├─Conv2d: 1-1                            [-1, 6, 24, 24]           156\n",
      "├─Conv2d: 1-2                            [-1, 16, 8, 8]            2,416\n",
      "├─Linear: 1-3                            [-1, 120]                 30,840\n",
      "├─Linear: 1-4                            [-1, 84]                  10,164\n",
      "├─Linear: 1-5                            [-1, 10]                  850\n",
      "==========================================================================================\n",
      "Total params: 44,426\n",
      "Trainable params: 44,426\n",
      "Non-trainable params: 0\n",
      "Total mult-adds (M): 0.28\n",
      "==========================================================================================\n",
      "Input size (MB): 0.00\n",
      "Forward/backward pass size (MB): 0.04\n",
      "Params size (MB): 0.17\n",
      "Estimated Total Size (MB): 0.21\n",
      "==========================================================================================\n"
     ]
    }
   ],
   "metadata": {}
  },
  {
   "cell_type": "markdown",
   "source": [
    "### 8.Loss Function and Optimizer"
   ],
   "metadata": {}
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "source": [
    "# Loss Function\r\n",
    "loss_fn = nn.CrossEntropyLoss()\r\n",
    "\r\n",
    "# Optimizer\r\n",
    "optimizer = optim.SGD(net.parameters(), lr=LEARNING_RATE, momentum=0.9)"
   ],
   "outputs": [],
   "metadata": {}
  },
  {
   "cell_type": "markdown",
   "source": [
    "### 9.Training"
   ],
   "metadata": {}
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "source": [
    "# function to calculate accuracy\r\n",
    "def calc_acc(loader):\r\n",
    "    correct = 0\r\n",
    "    total = 0\r\n",
    "    for data in loader:\r\n",
    "        inputs, labels = data[0].to(device), data[1].to(device)\r\n",
    "        outputs = net(inputs)\r\n",
    "        _, predicted = torch.max(outputs.data, 1)\r\n",
    "        total += labels.size(0)\r\n",
    "        correct += (predicted == labels).sum().item()\r\n",
    "    return ((100 * correct)/ total)\r\n"
   ],
   "outputs": [],
   "metadata": {}
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "source": [
    "def training():\r\n",
    "    epoch_loss = []\r\n",
    "    train_acc = []\r\n",
    "    test_acc = []\r\n",
    "    # 1. Number of epoches defined\r\n",
    "    for epoch in range(NUM_EPOCHS):\r\n",
    "        running_loss = 0.0\r\n",
    "        # 2. Iterate over data from trainloader\r\n",
    "        for i, data in enumerate(trainloader, 0):\r\n",
    "            # 3. Get inputs and labels from data\r\n",
    "            inputs, labels = data[0].to(device), data[1].to(device)\r\n",
    "            \r\n",
    "            # 4. Zero the gradient\r\n",
    "            optimizer.zero_grad()\r\n",
    "            \r\n",
    "            # 5. Forward pass\r\n",
    "            outputs = net(inputs)\r\n",
    "            \r\n",
    "            # 6. forward pass: compute loss\r\n",
    "            loss = loss_fn(outputs, labels)\r\n",
    "            \r\n",
    "            # 7. backward pass: compute gradient\r\n",
    "            loss.backward()\r\n",
    "            \r\n",
    "            # 8. update weights\r\n",
    "            optimizer.step()\r\n",
    "            \r\n",
    "            # 9. statistics\r\n",
    "            running_loss += loss.item()\r\n",
    "            \r\n",
    "        epoch_loss.append(running_loss)\r\n",
    "        train_acc.append(calc_acc(trainloader))\r\n",
    "        test_acc.append(calc_acc(testloader))\r\n",
    "        print('Epoch: %d of %d, Train Acc: %0.3f, Test Acc: %0.3f, Loss: %0.3f'\r\n",
    "              % (epoch+1, NUM_EPOCHS, train_acc[epoch], test_acc[epoch], running_loss/15000))\r\n",
    "        \r\n",
    "    return epoch_loss, train_acc, test_acc\r\n",
    "            \r\n",
    "    "
   ],
   "outputs": [],
   "metadata": {}
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "source": [
    "start = time.time()\r\n",
    "epoch_loss, train_acc, test_acc = training()\r\n",
    "end = time.time()\r\n",
    "\r\n",
    "print(\"%0.2f Minutes\" %((end-start)/60))"
   ],
   "outputs": [
    {
     "output_type": "stream",
     "name": "stdout",
     "text": [
      "Epoch: 1 of 10, Train Acc: 85.565, Test Acc: 84.210, Loss: 0.448\n",
      "Epoch: 2 of 10, Train Acc: 87.103, Test Acc: 85.790, Loss: 0.356\n",
      "Epoch: 3 of 10, Train Acc: 88.820, Test Acc: 87.430, Loss: 0.319\n",
      "Epoch: 4 of 10, Train Acc: 89.725, Test Acc: 88.050, Loss: 0.296\n",
      "Epoch: 5 of 10, Train Acc: 90.102, Test Acc: 87.970, Loss: 0.279\n",
      "Epoch: 6 of 10, Train Acc: 91.237, Test Acc: 89.110, Loss: 0.264\n",
      "Epoch: 7 of 10, Train Acc: 90.547, Test Acc: 88.400, Loss: 0.250\n",
      "Epoch: 8 of 10, Train Acc: 91.503, Test Acc: 88.840, Loss: 0.241\n",
      "Epoch: 9 of 10, Train Acc: 91.218, Test Acc: 88.730, Loss: 0.231\n",
      "Epoch: 10 of 10, Train Acc: 91.853, Test Acc: 88.790, Loss: 0.222\n",
      "14.76 Minutes\n"
     ]
    }
   ],
   "metadata": {}
  },
  {
   "cell_type": "markdown",
   "source": [
    "The loss is around 0.222 by the end of 10 epochs. Surely, it is not state of the art result, but it is not bad for such a simple network also. The training accuracy is above 91.853 % and the test accuracy is around 88.790 %. A bigger network will surely help in getting better results."
   ],
   "metadata": {}
  },
  {
   "cell_type": "markdown",
   "source": [
    "### 10.Plot the Loss and Accuracy"
   ],
   "metadata": {}
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "source": [
    "plt.figure(figsize=(20,6))\r\n",
    "plt.subplot(1,3,1)\r\n",
    "plt.plot(epoch_loss)\r\n",
    "plt.title(\"Loss\")\r\n",
    "plt.xlabel('Epoch')\r\n",
    "plt.ylabel('Loss')\r\n",
    "\r\n",
    "plt.subplot(1,3,2)\r\n",
    "plt.plot(train_acc)\r\n",
    "plt.title(\"Train Accuracy\")\r\n",
    "plt.xlabel('Epoch')\r\n",
    "plt.ylabel('Training Accuracy')\r\n",
    "\r\n",
    "plt.subplot(1,3,3)\r\n",
    "plt.plot(test_acc)\r\n",
    "plt.title(\"Testing Accuracy\")\r\n",
    "plt.xlabel('Epoch')\r\n",
    "plt.ylabel('Test Accuracy')\r\n",
    "plt.show()"
   ],
   "outputs": [
    {
     "output_type": "display_data",
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1440x432 with 3 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     }
    }
   ],
   "metadata": {}
  }
 ],
 "metadata": {
  "orig_nbformat": 4,
  "language_info": {
   "name": "python",
   "version": "3.8.0",
   "mimetype": "text/x-python",
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "pygments_lexer": "ipython3",
   "nbconvert_exporter": "python",
   "file_extension": ".py"
  },
  "kernelspec": {
   "name": "python3",
   "display_name": "Python 3.8.0 64-bit ('pytorch19': conda)"
  },
  "interpreter": {
   "hash": "884a8f81666a19c0851426c83cd6eaa7b212468ad852fb3caa21591c98d4369f"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}