Neural_Networks_MNIST_Classification.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "Neural Networks - MNIST Classification.ipynb",
      "provenance": [],
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/Meet-Patel2580/Reinforcement_Learning/blob/main/Neural_Networks_MNIST_Classification.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "cp7Xwf3A98SP"
      },
      "source": [
        "# MNIST DATASET\n",
        "\n",
        "convnets: \n",
        "https://ujjwalkarn.me/2016/08/11/intuitive-explanation-convnets/\n",
        "\n",
        "mnist 1\n",
        "https://towardsdatascience.com/handwritten-digit-mnist-pytorch-977b5338e627\n",
        "\n",
        "mnist 2 \n",
        "https://nextjournal.com/gkoehler/pytorch-mnist\n",
        "\n",
        "\n",
        "cifar10\n",
        "https://pytorch.org/tutorials/beginner/blitz/cifar10_tutorial.html\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ESuQ-wkl93lP"
      },
      "source": [
        "#imports\r\n",
        "import numpy as np\r\n",
        "import torch\r\n",
        "import torchvision\r\n",
        "import matplotlib.pyplot as plt\r\n",
        "from time import time\r\n",
        "from torch import nn, optim\r\n",
        "\r\n",
        "!wget www.di.ens.fr/~lelarge/MNIST.tar.gz\r\n",
        "!tar -zxvf MNIST.tar.gz\r\n",
        "\r\n",
        "from torchvision.datasets import MNIST\r\n",
        "from torchvision import transforms"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "lTUdI49wIztz"
      },
      "source": [
        "#dowloading images \r\n",
        "train_set = MNIST('./', download=True,\r\n",
        "transform=transforms.Compose([\r\n",
        "transforms.ToTensor(),\r\n",
        "]), train=True)\r\n",
        "\r\n",
        "test_set = MNIST('./', download=True,\r\n",
        "transform=transforms.Compose([\r\n",
        "transforms.ToTensor(),\r\n",
        "]), train=False)\r\n",
        "\r\n",
        "trainloader = torch.utils.data.DataLoader(train_set, batch_size=64, shuffle=True)\r\n",
        "valloader = torch.utils.data.DataLoader(test_set, batch_size=64, shuffle=True)"
      ],
      "execution_count": 40,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "iwh3HU6vIgJV",
        "outputId": "70d1526b-8a39-4542-f833-d91b87ce43ee"
      },
      "source": [
        "dataiter = iter(trainloader)\r\n",
        "images, labels = dataiter.next()\r\n",
        "\r\n",
        "#number of images: 64, dimensions: 28x28 pixels\r\n",
        "#64 labels\r\n",
        "print(images.shape)\r\n",
        "print(labels.shape)"
      ],
      "execution_count": 41,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "torch.Size([64, 1, 28, 28])\n",
            "torch.Size([64])\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 493
        },
        "id": "6veaBWfJ37al",
        "outputId": "66e23b8b-8097-4f20-9694-264dea9ae7ee"
      },
      "source": [
        "#displaying image\r\n",
        "plt.imshow(images[0].numpy().squeeze(), cmap='gray_r');\r\n",
        "\r\n",
        "#displaying data set\r\n",
        "figure = plt.figure()\r\n",
        "num_of_images = 60\r\n",
        "for index in range(1, num_of_images + 1):\r\n",
        "    plt.subplot(6, 10, index)\r\n",
        "    plt.axis('off')\r\n",
        "    plt.imshow(images[index].numpy().squeeze(), cmap='gray_r')"
      ],
      "execution_count": 42,
      "outputs": [
        {
          "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"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 60 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "3ICPDE8O6Vqq",
        "outputId": "034b8070-ac16-43f5-99e1-4b9b3879132f"
      },
      "source": [
        "#creating the neural network\r\n",
        "#input size is 784,  layer 1: 128, layer 2:64 and output consists of 10 units \r\n",
        "input_size = 784\r\n",
        "hidden_sizes = [128, 64]\r\n",
        "output_size = 10\r\n",
        "\r\n",
        "#wraps layers together, 3 layers with Relu activation which turns all negative values to 0\r\n",
        "model = nn.Sequential(nn.Linear(input_size, hidden_sizes[0]),\r\n",
        "                      nn.ReLU(),\r\n",
        "                      nn.Linear(hidden_sizes[0], hidden_sizes[1]),\r\n",
        "                      nn.ReLU(),\r\n",
        "                      nn.Linear(hidden_sizes[1], output_size),\r\n",
        "                      nn.LogSoftmax(dim=1))\r\n",
        "print(model)"
      ],
      "execution_count": 43,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Sequential(\n",
            "  (0): Linear(in_features=784, out_features=128, bias=True)\n",
            "  (1): ReLU()\n",
            "  (2): Linear(in_features=128, out_features=64, bias=True)\n",
            "  (3): ReLU()\n",
            "  (4): Linear(in_features=64, out_features=10, bias=True)\n",
            "  (5): LogSoftmax(dim=1)\n",
            ")\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "TIc2ZvH5NYdI"
      },
      "source": [
        "criterion = nn.NLLLoss()\r\n",
        "images, labels = next(iter(trainloader))\r\n",
        "images = images.view(images.shape[0], -1)\r\n",
        "\r\n",
        "logps = model(images) #log probabilities\r\n",
        "loss = criterion(logps, labels) #calculate the NLL loss"
      ],
      "execution_count": 44,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "UOOdIBS_OrIJ",
        "outputId": "4f908146-69af-4bd9-aace-90a01c067198"
      },
      "source": [
        "#training process\r\n",
        "\r\n",
        "optimizer = optim.SGD(model.parameters(), lr=0.003, momentum=0.9)\r\n",
        "time0 = time()\r\n",
        "epochs = 30\r\n",
        "for e in range(epochs):\r\n",
        "    running_loss = 0\r\n",
        "    for images, labels in trainloader:\r\n",
        "        # Flatten MNIST images into a 784 long vector\r\n",
        "        images = images.view(images.shape[0], -1)\r\n",
        "    \r\n",
        "        # Training pass\r\n",
        "        optimizer.zero_grad()\r\n",
        "        \r\n",
        "        output = model(images) #pass image through network\r\n",
        "        loss = criterion(output, labels) #calculate loss\r\n",
        "        \r\n",
        "        #This is where the model learns by backpropagating\r\n",
        "        loss.backward()\r\n",
        "        \r\n",
        "        #And optimizes its weights here\r\n",
        "        optimizer.step()\r\n",
        "        \r\n",
        "        running_loss += loss.item()\r\n",
        "    else:\r\n",
        "        print(\"Epoch {} - Training loss: {}\".format(e, running_loss/len(trainloader)))\r\n",
        "print(\"\\nTraining Time (in minutes) =\",(time()-time0)/60)"
      ],
      "execution_count": 45,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Epoch 0 - Training loss: 0.9588554034005604\n",
            "Epoch 1 - Training loss: 0.32894808720392205\n",
            "Epoch 2 - Training loss: 0.2724363042522214\n",
            "Epoch 3 - Training loss: 0.229724504184653\n",
            "Epoch 4 - Training loss: 0.19792081794139546\n",
            "Epoch 5 - Training loss: 0.17238421234558385\n",
            "Epoch 6 - Training loss: 0.15132373691335924\n",
            "Epoch 7 - Training loss: 0.13415326984492001\n",
            "Epoch 8 - Training loss: 0.12015619746316025\n",
            "Epoch 9 - Training loss: 0.10809949611418887\n",
            "Epoch 10 - Training loss: 0.09794974412375898\n",
            "Epoch 11 - Training loss: 0.0897954318350725\n",
            "Epoch 12 - Training loss: 0.08231386472719279\n",
            "Epoch 13 - Training loss: 0.07642122191776123\n",
            "Epoch 14 - Training loss: 0.07078618910372146\n",
            "Epoch 15 - Training loss: 0.06595551077105692\n",
            "Epoch 16 - Training loss: 0.061225342838816456\n",
            "Epoch 17 - Training loss: 0.05687514895340924\n",
            "Epoch 18 - Training loss: 0.053605240177927116\n",
            "Epoch 19 - Training loss: 0.05005481145702112\n",
            "Epoch 20 - Training loss: 0.04707198309824308\n",
            "Epoch 21 - Training loss: 0.04422516941722395\n",
            "Epoch 22 - Training loss: 0.04139240807853639\n",
            "Epoch 23 - Training loss: 0.03888060137662274\n",
            "Epoch 24 - Training loss: 0.03653018054381601\n",
            "Epoch 25 - Training loss: 0.034233912868650636\n",
            "Epoch 26 - Training loss: 0.0321257460320285\n",
            "Epoch 27 - Training loss: 0.03029420416489907\n",
            "Epoch 28 - Training loss: 0.028201090975546402\n",
            "Epoch 29 - Training loss: 0.02679797919662689\n",
            "\n",
            "Training Time (in minutes) = 2.6635277311007184\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "3ml9J7lrWnZ-"
      },
      "source": [
        "def view_classify(img, ps):\r\n",
        "    ''' Function for viewing an image and it's predicted classes.\r\n",
        "    '''\r\n",
        "    ps = ps.data.numpy().squeeze()\r\n",
        "\r\n",
        "    fig, (ax1, ax2) = plt.subplots(figsize=(6,9), ncols=2)\r\n",
        "    ax1.imshow(img.resize_(1, 28, 28).numpy().squeeze())\r\n",
        "    ax1.axis('off')\r\n",
        "    ax2.barh(np.arange(10), ps)\r\n",
        "    ax2.set_aspect(0.1)\r\n",
        "    ax2.set_yticks(np.arange(10))\r\n",
        "    ax2.set_yticklabels(np.arange(10))\r\n",
        "    ax2.set_title('Class Probability')\r\n",
        "    ax2.set_xlim(0, 1.1)\r\n",
        "    plt.tight_layout()"
      ],
      "execution_count": 46,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 268
        },
        "id": "xpKXyasqPDUK",
        "outputId": "807089b5-d046-4134-d5ea-9ffe5c792654"
      },
      "source": [
        "images, labels = next(iter(valloader))\r\n",
        "\r\n",
        "img = images[0].view(1, 784)\r\n",
        "with torch.no_grad():\r\n",
        "    logps = model(img)\r\n",
        "\r\n",
        "ps = torch.exp(logps)\r\n",
        "probab = list(ps.numpy()[0])\r\n",
        "print(\"Predicted Digit =\", probab.index(max(probab)))\r\n",
        "view_classify(img.view(1, 28, 28), ps)"
      ],
      "execution_count": 76,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Predicted Digit = 7\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x648 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "V32VDxEfZGvx",
        "outputId": "793f3cf0-fa3c-4e5a-f87b-c957bb7c92df"
      },
      "source": [
        "\r\n",
        "\r\n",
        "correct_count, all_count = 0, 0\r\n",
        "for images,labels in valloader:\r\n",
        "  for i in range(len(labels)):\r\n",
        "    img = images[i].view(1, 784)\r\n",
        "    with torch.no_grad():\r\n",
        "        logps = model(img)\r\n",
        "\r\n",
        "    \r\n",
        "    ps = torch.exp(logps)\r\n",
        "    probab = list(ps.numpy()[0])\r\n",
        "    pred_label = probab.index(max(probab))\r\n",
        "    true_label = labels.numpy()[i]\r\n",
        "    if(true_label == pred_label):\r\n",
        "      correct_count += 1\r\n",
        "    all_count += 1\r\n",
        "\r\n",
        "print(\"Number Of Images Tested =\", all_count)\r\n",
        "print(\"\\nModel Accuracy =\", (correct_count/all_count))"
      ],
      "execution_count": 48,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Number Of Images Tested = 10000\n",
            "\n",
            "Model Accuracy = 0.978\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "I1Px0hn7ZtzI"
      },
      "source": [
        "torch.save(model, './my_mnist_model.pt') "
      ],
      "execution_count": 67,
      "outputs": []
    }
  ]
}