exercise5-1.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Training CIFAR10 Classifier\n",
    "Follow the Sung Kim exercise by changing the `optimizers` not network!\n",
    "DPN network takes 7hrs with `p3.2xlarge`\n",
    "\n",
    "References:<br> \n",
    "https://www.youtube.com/watch?v=113b7O3mabY&list=PLlMkM4tgfjnJ3I-dbhO9JTw7gNty6o_2m&index=5<br>\n",
    "https://www.stefanfiott.com/machine-learning/cifar-10-classifier-using-cnn-in-pytorch/<br>\n",
    "git@github.com:kuangliu/pytorch-cifar.git"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [],
   "source": [
    "import torch\n",
    "import torch.nn as nn\n",
    "import torchvision\n",
    "import torchvision.transforms as transforms\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Define network\n",
    "class Net(nn.Module):\n",
    "    def __init__(self):\n",
    "        super().__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(nn.functional.relu(self.conv1(x)))\n",
    "        x = self.pool(nn.functional.relu(self.conv2(x)))\n",
    "        x = x.view(-1, 16 * 5 * 5)\n",
    "        x = nn.functional.relu(self.fc1(x))\n",
    "        x = nn.functional.relu(self.fc2(x))\n",
    "        x = self.fc3(x)\n",
    "        return x        "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Define a Loss function and optimizer\n",
    "net = Net()\n",
    "criterion = nn.CrossEntropyLoss()\n",
    "# optimizer = torch.optim.SGD(net.parameters(), lr=0.001, momentum=0.9)\n",
    "# optimizer = torch.optim.Rprop(net.parameters(), lr=0.001)\n",
    "optimizer = torch.optim.RMSprop(net.parameters(), lr=0.001)# best one\n",
    "# optimizer = torch.optim.ASGD(net.parameters(), lr=0.001)\n",
    "# optimizer = torch.optim.Adagrad(net.parameters(), lr=0.001)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Files already downloaded and verified\n",
      "Files already downloaded and verified\n"
     ]
    }
   ],
   "source": [
    "# Net model Sung Kim, Official docs\n",
    "transform = transforms.Compose(\n",
    "    [transforms.ToTensor(),\n",
    "     transforms.Normalize((0.5, 0.5, 0.5), (0.5, 0.5, 0.5))])\n",
    "\n",
    "trainset = torchvision.datasets.CIFAR10(root='./cifar-10-classifier/traing_data/', \n",
    "                                        train=True,\n",
    "                                        download=True, \n",
    "                                        transform=transform)\n",
    "trainloader = torch.utils.data.DataLoader(trainset, \n",
    "                                          batch_size=4,\n",
    "                                          shuffle=True)\n",
    "\n",
    "testset = torchvision.datasets.CIFAR10(root='./cifar-10-classifier/testing_data', \n",
    "                                       train=False,\n",
    "                                       download=True, \n",
    "                                       transform=transform)\n",
    "testloader = torch.utils.data.DataLoader(testset, \n",
    "                                         batch_size=4,\n",
    "                                         shuffle=False)\n",
    "\n",
    "classes = ('plane', 'car', 'bird', 'cat', 'deer', \n",
    "           'dog', 'frog', 'horse', 'ship', 'truck')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1, 2000, 1.8543944607377052\n",
      "1, 4000, 1.603652361229062\n",
      "1, 6000, 1.528085955247283\n",
      "1, 8000, 1.4681542510390282\n",
      "1, 10000, 1.46179886084795\n",
      "1, 12000, 1.4098220385462046\n",
      "2, 2000, 1.35464448081702\n",
      "2, 4000, 1.3510149261020123\n",
      "2, 6000, 1.3524481324665247\n",
      "2, 8000, 1.324898198839277\n",
      "2, 10000, 1.3212777344360949\n",
      "2, 12000, 1.3152193854302168\n",
      "Finished Training\n",
      "Saved model parameters to disk\n"
     ]
    }
   ],
   "source": [
    "# Train the network\n",
    "import os\n",
    "model_directory_path = '/home/jovyan/work/cifar-10-classifier/model/'\n",
    "model_path = model_directory_path + 'cifar-10-cnn-model.pt'\n",
    "\n",
    "if not os.path.exists(model_path):\n",
    "    os.makedirs(model_directory_path)\n",
    "\n",
    "if os.path.isfile(model_path):\n",
    "    net.load_state_dict(torch.load(model_path))\n",
    "    print(f\"Loaded model parameters from disk.\")\n",
    "else:\n",
    "    for epoch in range(2):\n",
    "        running_loss = 0\n",
    "        for i, data in enumerate(trainloader, 0):\n",
    "            inputs, labels = data\n",
    "\n",
    "            # Computer loss\n",
    "            outputs = net(inputs)\n",
    "            loss = criterion(outputs, labels)\n",
    "\n",
    "            # Zero gradients, perform a backward pass, and update the weights\n",
    "            optimizer.zero_grad()\n",
    "            loss.backward()\n",
    "            optimizer.step()\n",
    "            \n",
    "            running_loss += loss.item()\n",
    "            if i % 2000 == 1999:\n",
    "                print(f\"{epoch + 1}, {i+1}, {running_loss/2000}\")\n",
    "                running_loss = 0.0\n",
    "    print(f\"Finished Training\")\n",
    "    torch.save(net.state_dict(), model_path)\n",
    "    print(f\"Saved model parameters to disk\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "def convert_to_imshow_format(image):\n",
    "    # first convert back to [0,1] range from [-1,1] range\n",
    "    image = image / 2 + 0.5\n",
    "    image = image.numpy()\n",
    "    # convert from CHW to HWC\n",
    "    # from 3x32x32 to 32x32x3\n",
    "    return image.transpose(1,2,0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x180 with 4 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Show the images\n",
    "dataiter = iter(testloader)\n",
    "images, labels = dataiter.next()\n",
    "\n",
    "fig, axes = plt.subplots(1, len(images), figsize=(12,2.5))\n",
    "for idx, image in enumerate(images):\n",
    "    axes[idx].imshow(convert_to_imshow_format(image))\n",
    "    axes[idx].set_title(classes[labels[idx]])\n",
    "    axes[idx].set_xticks([])\n",
    "    axes[idx].set_yticks([])\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model accuracy on 10000 test images: 52.95%\n"
     ]
    }
   ],
   "source": [
    "# Test the network and construct the confusion matrix\n",
    "total_correct = 0\n",
    "total_images = 0\n",
    "confusion_matrix = np.zeros([10,10], int)\n",
    "with torch.no_grad():\n",
    "    for data in testloader:\n",
    "        images, labels = data\n",
    "        outputs = net(images)\n",
    "        _, predicted = torch.max(outputs.data, 1)\n",
    "        total_images += labels.size(0)\n",
    "        total_correct += (predicted == labels).sum().item()\n",
    "        for i, l in enumerate(labels):\n",
    "            confusion_matrix[l.item(), predicted[i].item()] += 1 \n",
    "\n",
    "model_accuracy = total_correct / total_images * 100\n",
    "print('Model accuracy on {0} test images: {1:.2f}%'.format(total_images, model_accuracy))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Category   - Accuracy\n",
      "plane      - 59.8\n",
      "car        - 71.9\n",
      "bird       - 54.0\n",
      "cat        - 26.9\n",
      "deer       - 32.9\n",
      "dog        - 41.9\n",
      "frog       - 72.7\n",
      "horse      - 57.3\n",
      "ship       - 63.4\n",
      "truck      - 48.7\n"
     ]
    }
   ],
   "source": [
    "print('{0:10s} - {1}'.format('Category','Accuracy'))\n",
    "for i, r in enumerate(confusion_matrix):\n",
    "    print('{0:10s} - {1:.1f}'.format(classes[i], r[i]/np.sum(r)*100))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "SGD<br>\n",
    "Category   - Accuracy<br>\n",
    "plane      - 43.1<br>\n",
    "car        - 51.4<br>\n",
    "bird       - 43.1<br>\n",
    "cat        - 50.2<br>\n",
    "deer       - 30.6<br>\n",
    "dog        - 33.0<br>\n",
    "frog       - 81.6<br>\n",
    "horse      - 51.4<br>\n",
    "ship       - 84.2<br>\n",
    "truck      - 63.4<br>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Rprop<br>\n",
    "Category   - Accuracy<br>\n",
    "plane      - 38.0<br>\n",
    "car        - 20.1<br>\n",
    "bird       - 51.9<br>\n",
    "cat        - 8.4<br>\n",
    "deer       - 0.4<br>\n",
    "dog        - 2.9<br>\n",
    "frog       - 18.5<br>\n",
    "horse      - 33.5<br>\n",
    "ship       - 15.4<br>\n",
    "truck      - 26.6<br>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "RMSprop<br>\n",
    "Category   - Accuracy<br>\n",
    "plane      - 64.0<br>\n",
    "car        - 64.0<br>\n",
    "bird       - 57.8<br>\n",
    "cat        - 25.3<br>\n",
    "deer       - 35.9<br>\n",
    "dog        - 56.1<br>\n",
    "frog       - 70.2<br>\n",
    "horse      - 53.6<br>\n",
    "ship       - 70.8<br>\n",
    "truck      - 54.7"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "ASGD<br>\n",
    "Category   - Accuracy<br>\n",
    "plane      - 46.8<br>\n",
    "car        - 32.4<br>\n",
    "bird       - 8.9<br>\n",
    "cat        - 24.2<br>\n",
    "deer       - 18.3<br>\n",
    "dog        - 17.0<br>\n",
    "frog       - 63.1<br>\n",
    "horse      - 53.8<br>\n",
    "ship       - 29.4<br>\n",
    "truck      - 57.6"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Adamax<br>\n",
    "Category   - Accuracy<br>\n",
    "plane      - 57.6<br>\n",
    "car        - 78.2<br>\n",
    "bird       - 35.1<br>\n",
    "cat        - 36.2<br>\n",
    "deer       - 38.4<br>\n",
    "dog        - 37.8<br>\n",
    "frog       - 68.8<br>\n",
    "horse      - 71.6<br>\n",
    "ship       - 58.8<br>\n",
    "truck      - 58.4"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Adagrad<br>\n",
    "plane      - 49.5<br>\n",
    "car        - 36.8<br>\n",
    "bird       - 13.4<br>\n",
    "cat        - 16.5<br>\n",
    "deer       - 20.4<br>\n",
    "dog        - 41.9<br>\n",
    "frog       - 59.8<br>\n",
    "horse      - 37.6<br>\n",
    "ship       - 45.6<br>\n",
    "truck      - 51.4"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots(1,1,figsize=(8,6))\n",
    "ax.matshow(confusion_matrix, aspect='auto', vmin=0, vmax=1000, cmap=plt.get_cmap('Blues'))\n",
    "plt.ylabel('Actual Category')\n",
    "plt.yticks(range(10), classes)\n",
    "plt.xlabel('Predicted Category')\n",
    "plt.xticks(range(10), classes)\n",
    "plt.show()"
   ]
  },
  {
   "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": "3.6.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}