CNN.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "import torch\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "import torch.nn.functional as F\n",
    "from torch import nn\n",
    "from torchvision import datasets, transforms"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "device = torch.device(\"cuda:0\" if torch.cuda.is_available() else \"cpu\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "transform = transforms.Compose([transforms.Resize((28,28)),\n",
    "                               transforms.ToTensor(),\n",
    "                               transforms.Normalize((0.5,), (0.5,))\n",
    "                               ])\n",
    "training_dataset = datasets.MNIST(root='./data', train=True, download=True, transform=transform)\n",
    "validation_dataset = datasets.MNIST(root='./data', train=False, download=True, transform=transform)\n",
    "\n",
    "training_loader = torch.utils.data.DataLoader(training_dataset, batch_size=100, shuffle=True)\n",
    "validation_loader = torch.utils.data.DataLoader(validation_dataset, batch_size = 100, shuffle=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "def im_convert(tensor):\n",
    "  image = tensor.clone().detach().numpy()\n",
    "  image = image.transpose(1, 2, 0)\n",
    "  image = image * np.array((0.5, 0.5, 0.5)) + np.array((0.5, 0.5, 0.5))\n",
    "  image = image.clip(0, 1)\n",
    "  return image"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1800x288 with 20 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "dataiter = iter(training_loader)\n",
    "images, labels = dataiter.next()\n",
    "fig = plt.figure(figsize=(25, 4))\n",
    "\n",
    "for idx in np.arange(20):\n",
    "  ax = fig.add_subplot(2, 10, idx+1, xticks=[], yticks=[])\n",
    "  plt.imshow(im_convert(images[idx]))\n",
    "  ax.set_title([labels[idx].item()])\n",
    "  "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "class LeNet(nn.Module):\n",
    "    def __init__(self):\n",
    "      super().__init__()\n",
    "      self.conv1 = nn.Conv2d(1, 20, 5, 1)\n",
    "      self.conv2 = nn.Conv2d(20, 50, 5, 1)\n",
    "      self.fc1 = nn.Linear(4*4*50, 500)\n",
    "      self.dropout1 = nn.Dropout(0.5)\n",
    "      self.fc2 = nn.Linear(500, 10)\n",
    "    def forward(self, x):\n",
    "      x = F.relu(self.conv1(x))\n",
    "      x = F.max_pool2d(x, 2, 2)\n",
    "      x = F.relu(self.conv2(x))\n",
    "      x = F.max_pool2d(x, 2, 2)\n",
    "      x = x.view(-1, 4*4*50)\n",
    "      x = F.relu(self.fc1(x))\n",
    "      x = self.dropout1(x)\n",
    "      x = self.fc2(x)\n",
    "      return x"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "LeNet(\n",
       "  (conv1): Conv2d(1, 20, kernel_size=(5, 5), stride=(1, 1))\n",
       "  (conv2): Conv2d(20, 50, kernel_size=(5, 5), stride=(1, 1))\n",
       "  (fc1): Linear(in_features=800, out_features=500, bias=True)\n",
       "  (dropout1): Dropout(p=0.5, inplace=False)\n",
       "  (fc2): Linear(in_features=500, out_features=10, bias=True)\n",
       ")"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model = LeNet().to(device)\n",
    "model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "criterion = nn.CrossEntropyLoss()\n",
    "optimizer = torch.optim.Adam(model.parameters(), lr = 0.0001)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "epoch : 1\n",
      "training loss: 0.5812, acc 84.2400 \n",
      "validation loss: 0.1832, validation acc 94.6300 \n",
      "epoch : 2\n",
      "training loss: 0.1502, acc 95.5200 \n",
      "validation loss: 0.1042, validation acc 96.7900 \n",
      "epoch : 3\n",
      "training loss: 0.0999, acc 97.0367 \n",
      "validation loss: 0.0748, validation acc 97.5500 \n",
      "epoch : 4\n",
      "training loss: 0.0784, acc 97.6550 \n",
      "validation loss: 0.0600, validation acc 98.0800 \n",
      "epoch : 5\n",
      "training loss: 0.0648, acc 98.0500 \n",
      "validation loss: 0.0522, validation acc 98.2300 \n",
      "epoch : 6\n",
      "training loss: 0.0564, acc 98.2533 \n",
      "validation loss: 0.0451, validation acc 98.5000 \n",
      "epoch : 7\n",
      "training loss: 0.0497, acc 98.4900 \n",
      "validation loss: 0.0442, validation acc 98.5800 \n",
      "epoch : 8\n",
      "training loss: 0.0442, acc 98.6750 \n",
      "validation loss: 0.0421, validation acc 98.6900 \n",
      "epoch : 9\n",
      "training loss: 0.0409, acc 98.7667 \n",
      "validation loss: 0.0371, validation acc 98.8100 \n",
      "epoch : 10\n",
      "training loss: 0.0354, acc 98.8917 \n",
      "validation loss: 0.0397, validation acc 98.7700 \n",
      "epoch : 11\n",
      "training loss: 0.0346, acc 98.9900 \n",
      "validation loss: 0.0321, validation acc 98.9100 \n",
      "epoch : 12\n",
      "training loss: 0.0314, acc 99.0400 \n",
      "validation loss: 0.0334, validation acc 98.7600 \n",
      "epoch : 13\n",
      "training loss: 0.0294, acc 99.1150 \n",
      "validation loss: 0.0368, validation acc 98.8100 \n",
      "epoch : 14\n",
      "training loss: 0.0272, acc 99.1533 \n",
      "validation loss: 0.0293, validation acc 99.0800 \n",
      "epoch : 15\n",
      "training loss: 0.0245, acc 99.2167 \n",
      "validation loss: 0.0296, validation acc 98.9700 \n"
     ]
    }
   ],
   "source": [
    "epochs = 15\n",
    "running_loss_history = []\n",
    "running_corrects_history = []\n",
    "val_running_loss_history = []\n",
    "val_running_corrects_history = []\n",
    "\n",
    "for e in range(epochs):\n",
    "  \n",
    "  running_loss = 0.0\n",
    "  running_corrects = 0.0\n",
    "  val_running_loss = 0.0\n",
    "  val_running_corrects = 0.0\n",
    "  \n",
    "  for inputs, labels in training_loader:\n",
    "    inputs = inputs.to(device)\n",
    "    labels = labels.to(device)\n",
    "    outputs = model(inputs)\n",
    "    loss = criterion(outputs, labels)\n",
    "    \n",
    "    optimizer.zero_grad()\n",
    "    loss.backward()\n",
    "    optimizer.step()\n",
    "    \n",
    "    _, preds = torch.max(outputs, 1)\n",
    "    running_loss += loss.item()\n",
    "    running_corrects += torch.sum(preds == labels.data)\n",
    "\n",
    "  else:\n",
    "    with torch.no_grad():\n",
    "      for val_inputs, val_labels in validation_loader:\n",
    "        val_inputs = val_inputs.to(device)\n",
    "        val_labels = val_labels.to(device)\n",
    "        val_outputs = model(val_inputs)\n",
    "        val_loss = criterion(val_outputs, val_labels)\n",
    "        \n",
    "        _, val_preds = torch.max(val_outputs, 1)\n",
    "        val_running_loss += val_loss.item()\n",
    "        val_running_corrects += torch.sum(val_preds == val_labels.data)\n",
    "      \n",
    "    epoch_loss = running_loss/len(training_loader)\n",
    "    epoch_acc = running_corrects.float()/ len(training_loader)\n",
    "    running_loss_history.append(epoch_loss)\n",
    "    running_corrects_history.append(epoch_acc)\n",
    "    \n",
    "    val_epoch_loss = val_running_loss/len(validation_loader)\n",
    "    val_epoch_acc = val_running_corrects.float()/ len(validation_loader)\n",
    "    val_running_loss_history.append(val_epoch_loss)\n",
    "    val_running_corrects_history.append(val_epoch_acc)\n",
    "    print('epoch :', (e+1))\n",
    "    print('training loss: {:.4f}, acc {:.4f} '.format(epoch_loss, epoch_acc.item()))\n",
    "    print('validation loss: {:.4f}, validation acc {:.4f} '.format(val_epoch_loss, val_epoch_acc.item()))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x7fd2bd3e3280>"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(running_loss_history, label = 'training loss')\n",
    "plt.plot(val_running_loss_history, label = 'validation loss')\n",
    "plt.legend()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x7fd2bdc5bc10>"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(running_corrects_history, label = 'training accuracy')\n",
    "plt.plot(val_running_corrects_history, label = 'validation accuracy')\n",
    "plt.legend()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "[Image Link](https://images.homedepot-static.com/productImages/007164ea-d47e-4f66-8d8c-fd9f621984a2/svn/architectural-mailboxes-house-letters-numbers-3585b-5-64_1000.jpg) - These images will be used for further demonstations"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import requests\n",
    "from PIL import Image\n",
    "\n",
    "url = 'https://images.homedepot-static.com/productImages/007164ea-d47e-4f66-8d8c-fd9f621984a2/svn/architectural-mailboxes-house-letters-numbers-3585b-5-64_1000.jpg'\n",
    "response = requests.get(url, stream = True)\n",
    "img = Image.open(response.raw)\n",
    "plt.imshow(img)\n",
    "plt.grid()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "img = PIL.ImageOps.invert(img)\n",
    "img = img.convert('1')\n",
    "img = transform(img)\n",
    "plt.imshow(im_convert(img))\n",
    "plt.grid()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "5\n"
     ]
    }
   ],
   "source": [
    "images = img.to(device)\n",
    "image = images[0].unsqueeze(0).unsqueeze(0)\n",
    "output = model(image)\n",
    "_, pred = torch.max(output, 1)\n",
    "print(pred.item())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1800x288 with 20 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "dataiter = iter(validation_loader)\n",
    "images, labels = dataiter.next()\n",
    "images = images.to(device)\n",
    "labels = labels.to(device)\n",
    "output = model(images)\n",
    "_, preds = torch.max(output, 1)\n",
    "\n",
    "fig = plt.figure(figsize=(25, 4))\n",
    "\n",
    "for idx in np.arange(20):\n",
    "  ax = fig.add_subplot(2, 10, idx+1, xticks=[], yticks=[])\n",
    "  plt.imshow(im_convert(images[idx]))\n",
    "  ax.set_title(\"{} ({})\".format(str(preds[idx].item()), str(labels[idx].item())), color=(\"green\" if preds[idx]==labels[idx] else \"red\"))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "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": "3.8.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}