MNIST_with_MLP.ipynb

{
  "cells": [
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "Csg4Sfh4te18"
      },
      "outputs": [],
      "source": [
        "!pip install torch accelerate einops plotly"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 39,
      "metadata": {
        "id": "E_aXuL8EumAV"
      },
      "outputs": [],
      "source": [
        "%matplotlib inline\n",
        "\n",
        "import torch as t\n",
        "import torch.nn as nn\n",
        "import torch.nn.functional as F\n",
        "import numpy as np\n",
        "import matplotlib.pyplot as plt\n",
        "\n",
        "from torch.utils.data import DataLoader\n",
        "from torchvision import datasets\n",
        "from torchvision.transforms import ToTensor\n",
        "\n",
        "from torch_optimizers import SGD, Adam, RMSprop"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "Z6dmAKpI-bms"
      },
      "outputs": [],
      "source": [
        "batch_size = 32\n",
        "\n",
        "train_dataset = datasets.MNIST('./data', train=True, download=True, transform=ToTensor())\n",
        "test_dataset = datasets.MNIST('./data', train=False, transform=ToTensor())\n",
        "\n",
        "train_loader = DataLoader(dataset=train_dataset, batch_size=batch_size, shuffle=True)\n",
        "test_loader = DataLoader(dataset=test_dataset, batch_size=batch_size, shuffle=False)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 8,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "jWQKZOY6-mXK",
        "outputId": "a09d452e-c807-4e70-d177-80b56db3f527"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "data: torch.Size([32, 1, 28, 28]) type: torch.FloatTensor\n",
            "target: torch.Size([32]) type: torch.LongTensor\n"
          ]
        }
      ],
      "source": [
        "# For training with the MLP we will need to resize the input image to a vector\n",
        "\n",
        "for (data, target) in train_loader:\n",
        "    print('data:', data.size(), 'type:', data.type())\n",
        "    print('target:', target.size(), 'type:', target.type())\n",
        "    break"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 7,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 124
        },
        "id": "o3ItXQA5-vnH",
        "outputId": "dbb885c5-e55a-4a4f-8025-74867a6cecee"
      },
      "outputs": [
        {
          "data": {
            "image/png": "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",
            "text/plain": [
              "<Figure size 1000x100 with 10 Axes>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        }
      ],
      "source": [
        "pltsize=1\n",
        "plt.figure(figsize=(10*pltsize, pltsize))\n",
        "\n",
        "for i in range(10):\n",
        "    plt.subplot(1,10,i+1)\n",
        "    plt.axis('off')\n",
        "    plt.imshow(data[i,:,:,:].numpy().reshape(28,28), cmap=\"gray_r\")\n",
        "    plt.title('Class: '+str(target[i].item()))"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "a5eJnVhP_AXf"
      },
      "source": [
        "### Simple MLP"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 66,
      "metadata": {
        "id": "40yZU4Co-8p6"
      },
      "outputs": [],
      "source": [
        "class MLP(nn.Module):\n",
        "\n",
        "    def __init__(self, input_dim, hd1, num_classes):\n",
        "        super().__init__()\n",
        "\n",
        "        self.flatten = nn.Flatten()\n",
        "        self.relu = nn.ReLU()\n",
        "\n",
        "        self.fc1 = nn.Linear(input_dim, hd1)\n",
        "        self.fc2 = nn.Linear(hd1, num_classes)\n",
        "\n",
        "    def forward(self, x):\n",
        "\n",
        "        x = self.flatten(x)\n",
        "        x = self.relu(self.fc1(x))\n",
        "        return self.fc2(x)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 111,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Wduya_SjA0Zh",
        "outputId": "39794edd-a8be-407e-b4e5-b8718928da23"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Output Shape:  torch.Size([1, 10])\n"
          ]
        }
      ],
      "source": [
        "device = t.device(\"cuda\")\n",
        "\n",
        "model = MLP(28*28, 128, 10)\n",
        "model = model.to(device)\n",
        "\n",
        "input = t.rand(1, 28, 28, device=device)\n",
        "output = model(input)\n",
        "\n",
        "print(\"Output Shape: \", output.shape)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 112,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "dpHISjCLGnkY",
        "outputId": "f266c4d8-e7b3-4563-8841-f4a148bf5799"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "tensor([4], device='cuda:0')\n"
          ]
        }
      ],
      "source": [
        "y_pred = output.argmax(dim=1)\n",
        "\n",
        "print(y_pred)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 113,
      "metadata": {
        "id": "Mrwa4njABAME"
      },
      "outputs": [],
      "source": [
        "SGD_optim = SGD(model.parameters(), lr=2e-3, momentum=0.99)\n",
        "criterion = nn.CrossEntropyLoss()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 114,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Vgy1dNvKFvZD",
        "outputId": "1e5b1ba3-0591-46b2-f66c-02bbb778110f"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Num Batches:  1875\n",
            "Num Items:  60000\n"
          ]
        }
      ],
      "source": [
        "num_batches = len(train_loader)\n",
        "num_items = len(train_loader.dataset)\n",
        "\n",
        "print(\"Num Batches: \", num_batches)\n",
        "print(\"Num Items: \", num_items)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 115,
      "metadata": {
        "id": "zaXyeXR4FjoW"
      },
      "outputs": [],
      "source": [
        "def train(optimizer):\n",
        "    model.train()\n",
        "\n",
        "    total_loss = 0\n",
        "    total_correct = 0\n",
        "\n",
        "    for x, y in train_loader:\n",
        "        x = x.to(device)\n",
        "        y = y.to(device)\n",
        "\n",
        "        output = model(x)\n",
        "        y_pred = output.argmax(dim=1)\n",
        "\n",
        "        # Calculate the loss and backprop\n",
        "        loss = criterion(output, y)\n",
        "        total_loss += loss\n",
        "\n",
        "        loss.backward()\n",
        "        optimizer.step()\n",
        "        optimizer.zero_grad()\n",
        "\n",
        "        num_correct = (y_pred == y).bool().sum()\n",
        "        total_correct += num_correct.item()\n",
        "\n",
        "    train_loss = total_loss/num_batches\n",
        "    accuracy = total_correct/num_items\n",
        "    print(f\"Average loss: {train_loss:7f}, accuracy: {accuracy:.2%}\")\n",
        "\n",
        "    return train_loss.detach().item()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 116,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "hrPFtddCEQSz",
        "outputId": "2908878e-b58b-44af-d9c8-184df83cc01e"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Training epoch: 1\n",
            "Average loss: 0.375327, accuracy: 89.53%\n",
            "Training epoch: 2\n",
            "Average loss: 0.130427, accuracy: 96.19%\n",
            "Training epoch: 3\n",
            "Average loss: 0.089649, accuracy: 97.33%\n",
            "Training epoch: 4\n",
            "Average loss: 0.067820, accuracy: 97.89%\n",
            "Training epoch: 5\n",
            "Average loss: 0.055576, accuracy: 98.28%\n",
            "Training epoch: 6\n",
            "Average loss: 0.043798, accuracy: 98.63%\n",
            "Training epoch: 7\n",
            "Average loss: 0.034945, accuracy: 98.93%\n",
            "Training epoch: 8\n",
            "Average loss: 0.029166, accuracy: 99.12%\n",
            "Training epoch: 9\n",
            "Average loss: 0.025418, accuracy: 99.22%\n",
            "Training epoch: 10\n",
            "Average loss: 0.020798, accuracy: 99.42%\n",
            "CPU times: user 1min 28s, sys: 737 ms, total: 1min 29s\n",
            "Wall time: 1min 31s\n"
          ]
        }
      ],
      "source": [
        "%%time\n",
        "epochs = 10\n",
        "sgd_loss_values = []\n",
        "\n",
        "for epoch in range(epochs):\n",
        "    print(f\"Training epoch: {epoch+1}\")\n",
        "    epoch_loss = train(SGD_optim)\n",
        "    sgd_loss_values.append(epoch_loss)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "qtCasrRKJS7Y"
      },
      "source": [
        "#### Adam"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 118,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "neM3VnIYJL0W",
        "outputId": "29ed2f8d-95c1-4659-f204-985b03b25c45"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Output Shape:  torch.Size([1, 10])\n"
          ]
        }
      ],
      "source": [
        "device = t.device(\"cuda\")\n",
        "\n",
        "model = MLP(28*28, 128, 10)\n",
        "model = model.to(device)\n",
        "\n",
        "input = t.rand(1, 28, 28, device=device)\n",
        "output = model(input)\n",
        "\n",
        "print(\"Output Shape: \", output.shape)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 119,
      "metadata": {
        "id": "t8yzT54OJViS"
      },
      "outputs": [],
      "source": [
        "Adam_optim = Adam(model.parameters(), lr=2e-3, beta1= 0.9, beta2= 0.99)\n",
        "criterion = nn.CrossEntropyLoss()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 120,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "TkZ5kL9TJjNd",
        "outputId": "1c84dee4-4834-49df-abd6-42bac4a2fb27"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Training epoch: 1\n",
            "Average loss: 0.238595, accuracy: 92.98%\n",
            "Training epoch: 2\n",
            "Average loss: 0.100643, accuracy: 96.93%\n",
            "Training epoch: 3\n",
            "Average loss: 0.072457, accuracy: 97.74%\n",
            "Training epoch: 4\n",
            "Average loss: 0.055049, accuracy: 98.26%\n",
            "Training epoch: 5\n",
            "Average loss: 0.045976, accuracy: 98.53%\n",
            "Training epoch: 6\n",
            "Average loss: 0.037575, accuracy: 98.79%\n",
            "Training epoch: 7\n",
            "Average loss: 0.033008, accuracy: 98.86%\n",
            "Training epoch: 8\n",
            "Average loss: 0.026180, accuracy: 99.15%\n",
            "Training epoch: 9\n",
            "Average loss: 0.023334, accuracy: 99.24%\n",
            "Training epoch: 10\n",
            "Average loss: 0.020935, accuracy: 99.28%\n",
            "CPU times: user 1min 36s, sys: 767 ms, total: 1min 37s\n",
            "Wall time: 1min 38s\n"
          ]
        }
      ],
      "source": [
        "%%time\n",
        "epochs = 10\n",
        "adam_loss_values = []\n",
        "\n",
        "for epoch in range(epochs):\n",
        "    print(f\"Training epoch: {epoch+1}\")\n",
        "    epoch_loss = train(Adam_optim)\n",
        "    adam_loss_values.append(epoch_loss)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 121,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 409
        },
        "id": "Yp-cJ0l1Jm0Y",
        "outputId": "f1cc6b64-9a33-49ec-a856-5f8375d7aba9"
      },
      "outputs": [
        {
          "data": {
            "image/png": "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",
            "text/plain": [
              "<Figure size 600x400 with 1 Axes>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        }
      ],
      "source": [
        "plt.figure(figsize=(6, 4))\n",
        "\n",
        "plt.plot(range(1, epochs+1), adam_loss_values, label=\"Adam\")\n",
        "plt.plot(range(1, epochs+1), sgd_loss_values, label=\"SGD with Momentum\")\n",
        "plt.xlabel('Epoch')\n",
        "plt.ylabel('Loss')\n",
        "plt.title('Training Loss Curve')\n",
        "plt.legend()\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "OQMiACdmKOUD"
      },
      "outputs": [],
      "source": []
    }
  ],
  "metadata": {
    "accelerator": "GPU",
    "colab": {
      "gpuType": "T4",
      "provenance": []
    },
    "kernelspec": {
      "display_name": "Python 3",
      "name": "python3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "nbformat": 4,
  "nbformat_minor": 0
}