MNIST_CNNs.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "MNIST_CNNs.ipynb",
      "provenance": [],
      "collapsed_sections": [],
      "authorship_tag": "ABX9TyPOsl5FLe+58nFAWspOWwQz",
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/vishxm/DL-NBs/blob/main/MNIST_CNNs.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "q26gVa-tDh9P"
      },
      "source": [
        "# MNIST with CNNs"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "gl1C58PoDlbP"
      },
      "source": [
        "## Importing libraries"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "eBz-1mmQlJlu"
      },
      "source": [
        "import torch\n",
        "import torch.nn as nn\n",
        "import torch.nn.functional as F\n",
        "import torch.optim as optim\n",
        "from torchvision import datasets, transforms\n",
        "from torch.autograd import Variable\n",
        "import matplotlib.pyplot as plt\n",
        "import numpy as np\n",
        "import warnings\n",
        "warnings.filterwarnings(\"ignore\")"
      ],
      "execution_count": 16,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "num5AbkUDtrc"
      },
      "source": [
        "## Creating the model for our neural network"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "P0miShAQlaKs",
        "outputId": "2dd34c14-fb1c-4985-b37f-5d9c1524ff3a",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 121
        }
      },
      "source": [
        "class mnistModel(nn.Module): #inheriting from nn.Module class\n",
        "    def __init__(self):\n",
        "        super(mnistModel, self).__init__() \n",
        "        self.conv1 = nn.Conv2d(1, 32, 5, padding=2) ## in 1, out 32, with kernel size 5 and padding 2\n",
        "        self.conv2 = nn.Conv2d(32, 64, 5, padding=2) ## in 32, out 64, with kernel size 5 and padding 2\n",
        "        self.fc1 = nn.Linear(64*7*7, 1024) ## feature map size is 7*7 by pooling\n",
        "        self.fc2 = nn.Linear(1024, 10) \n",
        "\n",
        "    def forward(self, x):\n",
        "        x = F.max_pool2d(F.relu(self.conv1(x)), 2)\n",
        "        x = F.max_pool2d(F.relu(self.conv2(x)), 2)\n",
        "        x = x.view(-1, 64*7*7) #reshape\n",
        "        x = F.relu(self.fc1(x))\n",
        "        x = F.dropout(x, training=self.training)\n",
        "        x = self.fc2(x)\n",
        "        x = F.log_softmax(x)\n",
        "            \n",
        "        return x\n",
        "\n",
        "model = mnistModel()\n",
        "model.cuda()"
      ],
      "execution_count": 17,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "mnistModel(\n",
              "  (conv1): Conv2d(1, 32, kernel_size=(5, 5), stride=(1, 1), padding=(2, 2))\n",
              "  (conv2): Conv2d(32, 64, kernel_size=(5, 5), stride=(1, 1), padding=(2, 2))\n",
              "  (fc1): Linear(in_features=3136, out_features=1024, bias=True)\n",
              "  (fc2): Linear(in_features=1024, out_features=10, bias=True)\n",
              ")"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 17
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "oZ5rwIxYD014"
      },
      "source": [
        "## Defining train and test data loader"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "LGyMkVC5lb_d"
      },
      "source": [
        "batch_size = 50\n",
        "train_loader = torch.utils.data.DataLoader(\n",
        "    datasets.MNIST('data', train=True, download=True, transform=transforms.ToTensor()),\n",
        "    batch_size=batch_size, shuffle=True\n",
        ")"
      ],
      "execution_count": 18,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "dwXh7EYild9I"
      },
      "source": [
        "test_loader = torch.utils.data.DataLoader(\n",
        "    datasets.MNIST('data', train=False, transform=transforms.ToTensor()),\n",
        "    batch_size=1000\n",
        ")"
      ],
      "execution_count": 19,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ZUJNkkfAlgfS",
        "outputId": "f1bff04a-e2fa-4834-e048-e65dc44f4c88",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 156
        }
      },
      "source": [
        "for p in model.parameters():\n",
        "    print(p.size())"
      ],
      "execution_count": 20,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "torch.Size([32, 1, 5, 5])\n",
            "torch.Size([32])\n",
            "torch.Size([64, 32, 5, 5])\n",
            "torch.Size([64])\n",
            "torch.Size([1024, 3136])\n",
            "torch.Size([1024])\n",
            "torch.Size([10, 1024])\n",
            "torch.Size([10])\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "yazfHlOVliKH"
      },
      "source": [
        "optimizer = optim.Adam(model.parameters(), lr=0.0001)"
      ],
      "execution_count": 21,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Wzb3G1ljD8or"
      },
      "source": [
        "## Training the model with 15 epochs"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "bjw4E8K7ljma",
        "outputId": "c7c52552-b870-449e-9c12-d0afd229af5a",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 330
        }
      },
      "source": [
        "model.train()\n",
        "train_loss = []\n",
        "train_accu = []\n",
        "i = 0\n",
        "\n",
        "for epoch in range(15):\n",
        "    for data, target in train_loader:\n",
        "        data, target = Variable(data), Variable(target)\n",
        "        data, target = data.cuda(), target.cuda()\n",
        "        optimizer.zero_grad()\n",
        "        output = model(data)\n",
        "        loss = F.nll_loss(output, target)\n",
        "        loss.backward()\n",
        "        train_loss.append(loss.data)\n",
        "        optimizer.step()\n",
        "        prediction = output.data.max(1)[1]\n",
        "        accuracy = torch.true_divide(prediction.eq(target.data).sum() * 100,\n",
        "                            batch_size)\n",
        "        train_accu.append(accuracy)\n",
        "\n",
        "        if i%1000 == 0:\n",
        "            print(\"Train Step: {}\\tLoss: {:.3f}\\tAccuracy: {:.3f}\".format(i, loss.data, accuracy))\n",
        "\n",
        "        i += 1"
      ],
      "execution_count": 22,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Train Step: 0\tLoss: 2.314\tAccuracy: 6.000\n",
            "Train Step: 1000\tLoss: 0.121\tAccuracy: 96.000\n",
            "Train Step: 2000\tLoss: 0.145\tAccuracy: 92.000\n",
            "Train Step: 3000\tLoss: 0.007\tAccuracy: 100.000\n",
            "Train Step: 4000\tLoss: 0.032\tAccuracy: 100.000\n",
            "Train Step: 5000\tLoss: 0.063\tAccuracy: 96.000\n",
            "Train Step: 6000\tLoss: 0.095\tAccuracy: 96.000\n",
            "Train Step: 7000\tLoss: 0.066\tAccuracy: 96.000\n",
            "Train Step: 8000\tLoss: 0.029\tAccuracy: 98.000\n",
            "Train Step: 9000\tLoss: 0.008\tAccuracy: 100.000\n",
            "Train Step: 10000\tLoss: 0.004\tAccuracy: 100.000\n",
            "Train Step: 11000\tLoss: 0.006\tAccuracy: 100.000\n",
            "Train Step: 12000\tLoss: 0.007\tAccuracy: 100.000\n",
            "Train Step: 13000\tLoss: 0.005\tAccuracy: 100.000\n",
            "Train Step: 14000\tLoss: 0.002\tAccuracy: 100.000\n",
            "Train Step: 15000\tLoss: 0.031\tAccuracy: 98.000\n",
            "Train Step: 16000\tLoss: 0.003\tAccuracy: 100.000\n",
            "Train Step: 17000\tLoss: 0.001\tAccuracy: 100.000\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "33v0hsy7llAh",
        "outputId": "0ebbda6f-acb4-4c1e-82ba-7706f488edf4",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "torch.cuda.is_available()"
      ],
      "execution_count": 23,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "True"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 23
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "12Whwy6znGWl",
        "outputId": "0d9f1e6c-bf26-45d7-c3f2-62ee1d7b7482",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        }
      },
      "source": [
        "torch.cuda.get_device_name() #checking the GPU provided by Google Colab"
      ],
      "execution_count": 24,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "application/vnd.google.colaboratory.intrinsic+json": {
              "type": "string"
            },
            "text/plain": [
              "'Tesla P100-PCIE-16GB'"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 24
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "uZ5Wb1ASnI03",
        "outputId": "47eb7f19-902f-4155-82db-9ac039a2f4a2",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 282
        }
      },
      "source": [
        "plt.plot(np.arange(len(train_loss)), train_loss)"
      ],
      "execution_count": 25,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[<matplotlib.lines.Line2D at 0x7f10c1881940>]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 25
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "g1MS0RmOol2k",
        "outputId": "361d38b2-9d24-4945-d2ba-c7bc78d49e3a",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 282
        }
      },
      "source": [
        "plt.plot(np.arange(len(train_accu)), train_accu)"
      ],
      "execution_count": 26,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[<matplotlib.lines.Line2D at 0x7f10c198d2b0>]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 26
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "W53-IEggEGzM"
      },
      "source": [
        "## Evaluating the trained model"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "_3z3pZDWo3-Q",
        "outputId": "67ad3c3a-dacf-4e52-a879-5f92aed7f742",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 69
        }
      },
      "source": [
        "model.eval()\n",
        "correct = 0\n",
        "\n",
        "for data, target in test_loader:\n",
        "    data, target = Variable(data, volatile=True).cuda(), Variable(target).cuda()\n",
        "    output = model(data)\n",
        "    prediction = output.data.max(1)[1]\n",
        "    correct += prediction.eq(target.data).sum()\n",
        "\n",
        "print(\"\\n\\nTest set Accuracy : {:.2f}%\".format(100. * correct / len(test_loader.dataset)))"
      ],
      "execution_count": 27,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "\n",
            "\n",
            "Test set Accuracy : 99.23%\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "wtBRR7ippfg-"
      },
      "source": [
        ""
      ],
      "execution_count": 27,
      "outputs": []
    }
  ]
}