TorchLinMNIST.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "provenance": [],
      "authorship_tag": "ABX9TyPjfpj6pgW9pl3fOiSy0EiT",
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/diro5t/deep_quadric_learning/blob/main/TorchLinMNIST.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "(C) Dirk Roeckmann 2023\n",
        "\n",
        "Simple PyTorch MNIST ~100K parameter model classifier with linear layers: 784-128-10, batch size of 128 and 20 epochs.\n",
        "\n",
        "NOTE: The goal is NOT to find the minimum model or a certain target accuracy, but to compare it with a ~25K quadric model which achieves comparable accuracy after the same number of epochs (see TorchQuadMNIST.ipynb).\n",
        "\n",
        "More about quadric models here:\n",
        "\n",
        "https://github.com/diro5t/deep_quadric_learning\n",
        "\n",
        "This paradigm is based on my research here:\n",
        "\n",
        "https://scholar.google.com/citations?hl=en&user=hamuGh4AAAAJ"
      ],
      "metadata": {
        "id": "rHa0nm90z4-G"
      }
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "p3ecKA4tb0Fd"
      },
      "outputs": [],
      "source": [
        "import numpy as np\n",
        "import torch\n",
        "import torchvision\n",
        "import matplotlib.pyplot as plt\n",
        "from torchvision import datasets, transforms\n",
        "from torch import nn, optim"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "transf = transforms.Compose([transforms.ToTensor(),\n",
        "  transforms.Normalize((0.5,), (0.5,)),])"
      ],
      "metadata": {
        "id": "W7DZNjRucCUt"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "batch_size = 128\n",
        "\n",
        "train_set = datasets.MNIST('PATH_TO_STORE_TRAINSET', download=True, train=True, transform=transf)\n",
        "test_set = datasets.MNIST('PATH_TO_STORE_TESTSET', download=True, train=False, transform=transf)\n",
        "train_loader = torch.utils.data.DataLoader(train_set, batch_size=batch_size, shuffle=True)\n",
        "test_loader = torch.utils.data.DataLoader(test_set, batch_size=batch_size, shuffle=True)"
      ],
      "metadata": {
        "id": "eyEqc5stcWai"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "img_iter = iter(train_loader)\n",
        "images, labels = next(img_iter)"
      ],
      "metadata": {
        "id": "Sn3WjSKZk3uR"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "figure = plt.figure()\n",
        "img_num = 60\n",
        "for index in range(1, img_num + 1):\n",
        "    plt.subplot(10, 6, index)\n",
        "    plt.axis('off')\n",
        "    plt.imshow(images[index].numpy().squeeze(), cmap='gray_r')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 406
        },
        "id": "qfpOm6YxkjAo",
        "outputId": "027d6f16-2422-40be-8300-f31b4a001361"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 60 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "model = nn.Sequential(nn.Linear(784, 128),\n",
        "                      nn.ReLU(),\n",
        "                      nn.Linear(128, 10),\n",
        "                      nn.LogSoftmax(dim=1))"
      ],
      "metadata": {
        "id": "SAOnVadKn1ZY"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# total number of model parameters\n",
        "sum(p.numel() for p in model.parameters())"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "kEwSnq7Kdhvu",
        "outputId": "f3c8c040-9045-46c1-a497-addb850d70ae"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "101770"
            ]
          },
          "metadata": {},
          "execution_count": 154
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "loss_fn = nn.NLLLoss()\n",
        "optimizer = optim.SGD(model.parameters(), lr=0.003, momentum=0.9)\n",
        "epochs = 20\n",
        "for e in range(epochs):\n",
        "    train_loss = 0\n",
        "    for images, labels in train_loader:\n",
        "        images = images.view(images.shape[0], -1)\n",
        "        optimizer.zero_grad()\n",
        "        output = model(images)\n",
        "        loss = loss_fn(output, labels)\n",
        "        loss.backward()\n",
        "        optimizer.step()\n",
        "        train_loss += loss.item()\n",
        "    else:\n",
        "        print(\"Epoch {} - Training loss: {}\".format(e, train_loss/len(train_loader)))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "xfp1npFtpWZK",
        "outputId": "ffc66ed7-d12e-453f-bd8d-2cd78653b2fa"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 0 - Training loss: 0.6781163818038094\n",
            "Epoch 1 - Training loss: 0.33331811793450355\n",
            "Epoch 2 - Training loss: 0.2894533862754988\n",
            "Epoch 3 - Training loss: 0.2582498226147979\n",
            "Epoch 4 - Training loss: 0.23217386137574975\n",
            "Epoch 5 - Training loss: 0.21112761065077934\n",
            "Epoch 6 - Training loss: 0.19249198318862204\n",
            "Epoch 7 - Training loss: 0.1767189907335015\n",
            "Epoch 8 - Training loss: 0.16325940579366582\n",
            "Epoch 9 - Training loss: 0.15136547358845598\n",
            "Epoch 10 - Training loss: 0.1415880021732499\n",
            "Epoch 11 - Training loss: 0.13325830972366243\n",
            "Epoch 12 - Training loss: 0.12466952244419532\n",
            "Epoch 13 - Training loss: 0.11880632279428846\n",
            "Epoch 14 - Training loss: 0.11216575473245145\n",
            "Epoch 15 - Training loss: 0.1068294606308566\n",
            "Epoch 16 - Training loss: 0.10135210018708253\n",
            "Epoch 17 - Training loss: 0.09712600206801378\n",
            "Epoch 18 - Training loss: 0.09263223558584892\n",
            "Epoch 19 - Training loss: 0.08872691561370644\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "corrects, all = 0, 0\n",
        "for images,labels in train_loader:\n",
        "  for i in range(len(labels)):\n",
        "    img = images[i].view(1, 784)\n",
        "    with torch.inference_mode():\n",
        "        logps = model(img)\n",
        "    ps = torch.exp(logps)\n",
        "    prob = list(ps.numpy()[0])\n",
        "    inf_label = prob.index(max(prob))\n",
        "    true_label = labels.numpy()[i]\n",
        "    if(true_label == inf_label):\n",
        "      corrects += 1\n",
        "    all += 1\n",
        "\n",
        "print(\"Number of trained images=\", all)\n",
        "print(\"\\nModel Training Accuracy =\", (corrects / all))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "drZ4EhpftQPj",
        "outputId": "c6a923dd-f89d-4194-b2df-38da3dbff62f"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Number of trained images= 60000\n",
            "\n",
            "Model Training Accuracy = 0.9771\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "corrects, all = 0, 0\n",
        "for images,labels in test_loader:\n",
        "  for i in range(len(labels)):\n",
        "    img = images[i].view(1, 784)\n",
        "    with torch.inference_mode():\n",
        "        logps = model(img)\n",
        "    ps = torch.exp(logps)\n",
        "    prob = list(ps.numpy()[0])\n",
        "    inf_label = prob.index(max(prob))\n",
        "    true_label = labels.numpy()[i]\n",
        "    if(true_label == inf_label):\n",
        "      corrects += 1\n",
        "    all += 1\n",
        "\n",
        "print(\"Number of tested images=\", all)\n",
        "print(\"\\nModel Test Accuracy =\", (corrects / all))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Z3Far_XyUQxA",
        "outputId": "44d9aaf7-bab9-4f4b-c9a3-5e4c3ab4e5c2"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Number of tested images= 10000\n",
            "\n",
            "Model Test Accuracy = 0.9716\n"
          ]
        }
      ]
    }
  ]
}