Pytorch_Autoencoder_MNIST.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "provenance": [],
      "gpuType": "T4"
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "source": [
        "# Importing required python libraries"
      ],
      "metadata": {
        "id": "su7Yduixc0tF"
      }
    },
    {
      "cell_type": "code",
      "execution_count": 1,
      "metadata": {
        "id": "DNuUTKzLcmB0"
      },
      "outputs": [],
      "source": [
        "import torch\n",
        "import torchvision\n",
        "import torchvision.datasets as datasets\n",
        "import matplotlib.pyplot as plt\n",
        "import torch.nn as nn\n",
        "import torch.nn.functional as F\n",
        "import torch.optim as optim"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Downloading MNIST dataset"
      ],
      "metadata": {
        "id": "UG-rQpkRdyzX"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "batch_size_train = 64\n",
        "batch_size_test = 1000"
      ],
      "metadata": {
        "id": "AdLOSpWAdokS"
      },
      "execution_count": 2,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "training_data = torchvision.datasets.MNIST('/files/', train=True, download=True,\n",
        "                             transform=torchvision.transforms.Compose([\n",
        "                               torchvision.transforms.ToTensor(),\n",
        "                               torchvision.transforms.Normalize(\n",
        "                                 (0.1307,), (0.3081,))\n",
        "                             ]))\n",
        "test_data=torchvision.datasets.MNIST('/files/', train=False, download=True,\n",
        "                             transform=torchvision.transforms.Compose([\n",
        "                               torchvision.transforms.ToTensor(),\n",
        "                               torchvision.transforms.Normalize(\n",
        "                                 (0.1307,), (0.3081,))\n",
        "                             ]))"
      ],
      "metadata": {
        "id": "0mFK9u6kqWmj",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "bd06fc16-e19b-42e9-b666-3d521f6b1c88"
      },
      "execution_count": 3,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Downloading http://yann.lecun.com/exdb/mnist/train-images-idx3-ubyte.gz\n",
            "Downloading http://yann.lecun.com/exdb/mnist/train-images-idx3-ubyte.gz to /files/MNIST/raw/train-images-idx3-ubyte.gz\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "100%|██████████| 9912422/9912422 [00:00<00:00, 91478731.41it/s]\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Extracting /files/MNIST/raw/train-images-idx3-ubyte.gz to /files/MNIST/raw\n",
            "\n",
            "Downloading http://yann.lecun.com/exdb/mnist/train-labels-idx1-ubyte.gz\n",
            "Downloading http://yann.lecun.com/exdb/mnist/train-labels-idx1-ubyte.gz to /files/MNIST/raw/train-labels-idx1-ubyte.gz\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "100%|██████████| 28881/28881 [00:00<00:00, 113956438.22it/s]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Extracting /files/MNIST/raw/train-labels-idx1-ubyte.gz to /files/MNIST/raw\n",
            "\n",
            "Downloading http://yann.lecun.com/exdb/mnist/t10k-images-idx3-ubyte.gz\n",
            "Downloading http://yann.lecun.com/exdb/mnist/t10k-images-idx3-ubyte.gz to /files/MNIST/raw/t10k-images-idx3-ubyte.gz\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "\n",
            "100%|██████████| 1648877/1648877 [00:00<00:00, 42778092.25it/s]\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Extracting /files/MNIST/raw/t10k-images-idx3-ubyte.gz to /files/MNIST/raw\n",
            "\n",
            "Downloading http://yann.lecun.com/exdb/mnist/t10k-labels-idx1-ubyte.gz\n",
            "Downloading http://yann.lecun.com/exdb/mnist/t10k-labels-idx1-ubyte.gz to /files/MNIST/raw/t10k-labels-idx1-ubyte.gz\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "100%|██████████| 4542/4542 [00:00<00:00, 18640439.11it/s]\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Extracting /files/MNIST/raw/t10k-labels-idx1-ubyte.gz to /files/MNIST/raw\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "train_loader = torch.utils.data.DataLoader(training_data\n",
        "  ,\n",
        "  batch_size=batch_size_train, shuffle=True)\n",
        "\n",
        "test_loader = torch.utils.data.DataLoader(test_data\n",
        "  ,\n",
        "  batch_size=batch_size_test, shuffle=True)"
      ],
      "metadata": {
        "id": "4eef-PHOfVnz"
      },
      "execution_count": 36,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Data Visualisation\n"
      ],
      "metadata": {
        "id": "7_BvqKritGb2"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "labels_map = {\n",
        "    0: \"0\",\n",
        "    1: \"1\",\n",
        "    2: \"2\",\n",
        "    3: \"3\",\n",
        "    4: \"4\",\n",
        "    5: \"5\",\n",
        "    6: \"6\",\n",
        "    7: \"7\",\n",
        "    8: \"8\",\n",
        "    9: \"9\",\n",
        "}\n",
        "figure = plt.figure(figsize=(8, 8))\n",
        "cols, rows = 3, 3\n",
        "for i in range(1, cols * rows + 1):\n",
        "    sample_idx = torch.randint(len(training_data), size=(1,)).item()\n",
        "    img, label = training_data[sample_idx]\n",
        "    figure.add_subplot(rows, cols, i)\n",
        "    plt.title(labels_map[label])\n",
        "    plt.axis(\"off\")\n",
        "    plt.imshow(img.squeeze(), cmap=\"gray\")\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 675
        },
        "id": "YaI_fd07p-oj",
        "outputId": "3762d15a-fb64-4716-998b-cc99c25e770c"
      },
      "execution_count": 5,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 800x800 with 9 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "sample_idx = torch.randint(len(training_data), size=(1,)).item()\n",
        "img, label = training_data[sample_idx]"
      ],
      "metadata": {
        "id": "-HdHzJh0q7ys"
      },
      "execution_count": 6,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "img.shape"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "YONugKLXsRxC",
        "outputId": "3d3285ed-ac4e-4ce1-f770-f40a27f8a862"
      },
      "execution_count": 7,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "torch.Size([1, 28, 28])"
            ]
          },
          "metadata": {},
          "execution_count": 7
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Defining Autoencoder Architecture\n",
        "\n",
        "![autoencoder.png](data:image/png;base64,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)"
      ],
      "metadata": {
        "id": "TJuF91rEtYBP"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "class autoencoder(nn.Module):\n",
        "    def __init__(self):\n",
        "        super(autoencoder, self).__init__()\n",
        "        ## encoder layers ##\n",
        "\n",
        "        self.conv1 = nn.Conv2d(1, 32, 2, stride=2) #[(W−K)/S]+1   ((28-2/2)+1)\n",
        "        self.dropout = nn.Dropout(p=0.1)\n",
        "\n",
        "        self.conv2 = nn.Conv2d(32, 64, 2, stride=2) #((14-2/2)+1)\n",
        "        self.conv3 = nn.Conv2d(64, 128, 3, stride=2)#((7-3/2)+1)\n",
        "        self.flatten = nn.Flatten()\n",
        "        self.bottleneck=nn.Linear(3*3*128,10)\n",
        "        self.decoderlinear=nn.Linear(10,3*3*128)\n",
        "        ## decoder layers ##\n",
        "        ## a kernel of 2 and a stride of 2 will increase the spatial dims by 2\n",
        "        self.t_conv1 = nn.ConvTranspose2d(128, 64, 3, stride=2)\n",
        "        self.t_conv2 = nn.ConvTranspose2d(64, 32, 2, stride=2)\n",
        "        self.t_conv3 = nn.ConvTranspose2d(32, 1, 2, stride=2)\n",
        "    def forward(self, x):\n",
        "        ## encode ##\n",
        "        # add hidden layers with relu activation function\n",
        "        x = F.relu(self.conv1(x))\n",
        "        x = self.dropout(x)\n",
        "        x = F.relu(self.conv2(x))\n",
        "        x = self.dropout(x)\n",
        "        x = F.relu(self.conv3(x))\n",
        "        x = self.dropout(x)\n",
        "        x = self.flatten(x)\n",
        "        x1 = F.softmax(self.bottleneck(x))\n",
        "        x = F.softmax(self.bottleneck(x))\n",
        "        x = self.decoderlinear(x1)\n",
        "        x = x.view(x.size(0), 128, 3, 3)\n",
        "\n",
        "        ## decode ##\n",
        "        # add transpose conv layers, with relu activation function\n",
        "        x = self.t_conv1(x)\n",
        "        x = self.dropout(x)\n",
        "        x = self.t_conv2(x)\n",
        "        x = self.dropout(x)\n",
        "        # output layer (with sigmoid for scaling from 0 to 1)\n",
        "        x = F.sigmoid(self.t_conv3(x))\n",
        "        return x , x1\n",
        "# initialize the NN\n",
        "model = autoencoder()"
      ],
      "metadata": {
        "id": "dONTnYxLsStt"
      },
      "execution_count": 58,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "print(model)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "tXzzdvPC5EKa",
        "outputId": "a61b43c6-c12d-4773-9246-6857429b25b0"
      },
      "execution_count": 59,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "autoencoder(\n",
            "  (conv1): Conv2d(1, 32, kernel_size=(2, 2), stride=(2, 2))\n",
            "  (dropout): Dropout(p=0.1, inplace=False)\n",
            "  (conv2): Conv2d(32, 64, kernel_size=(2, 2), stride=(2, 2))\n",
            "  (conv3): Conv2d(64, 128, kernel_size=(3, 3), stride=(2, 2))\n",
            "  (flatten): Flatten(start_dim=1, end_dim=-1)\n",
            "  (bottleneck): Linear(in_features=1152, out_features=10, bias=True)\n",
            "  (decoderlinear): Linear(in_features=10, out_features=1152, bias=True)\n",
            "  (t_conv1): ConvTranspose2d(128, 64, kernel_size=(3, 3), stride=(2, 2))\n",
            "  (t_conv2): ConvTranspose2d(64, 32, kernel_size=(2, 2), stride=(2, 2))\n",
            "  (t_conv3): ConvTranspose2d(32, 1, kernel_size=(2, 2), stride=(2, 2))\n",
            ")\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "#Loss function\n",
        "#criterion = nn.MSELoss()\n",
        "distance   = nn.MSELoss()\n",
        "class_loss = nn.CrossEntropyLoss()\n",
        "\n",
        "#optimizer = torch.optim.SGD(model.parameters(), lr=0.01, momentum=0.5)\n",
        "#Optimizer\n",
        "optimizer = torch.optim.Adam(model.parameters(), lr=0.001)"
      ],
      "metadata": {
        "id": "19-x4kLv5Rxb"
      },
      "execution_count": 60,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "def get_device():\n",
        "    if torch.cuda.is_available():\n",
        "        device = 'cuda:0'\n",
        "    else:\n",
        "        device = 'cpu'\n",
        "    return device\n",
        "\n",
        "device = get_device()\n",
        "print(device)\n",
        "model.to(device)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "6U11kuH-5zqu",
        "outputId": "41729fad-5360-463c-d994-c486b4bcd3dd"
      },
      "execution_count": 61,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "cuda:0\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "autoencoder(\n",
              "  (conv1): Conv2d(1, 32, kernel_size=(2, 2), stride=(2, 2))\n",
              "  (dropout): Dropout(p=0.1, inplace=False)\n",
              "  (conv2): Conv2d(32, 64, kernel_size=(2, 2), stride=(2, 2))\n",
              "  (conv3): Conv2d(64, 128, kernel_size=(3, 3), stride=(2, 2))\n",
              "  (flatten): Flatten(start_dim=1, end_dim=-1)\n",
              "  (bottleneck): Linear(in_features=1152, out_features=10, bias=True)\n",
              "  (decoderlinear): Linear(in_features=10, out_features=1152, bias=True)\n",
              "  (t_conv1): ConvTranspose2d(128, 64, kernel_size=(3, 3), stride=(2, 2))\n",
              "  (t_conv2): ConvTranspose2d(64, 32, kernel_size=(2, 2), stride=(2, 2))\n",
              "  (t_conv3): ConvTranspose2d(32, 1, kernel_size=(2, 2), stride=(2, 2))\n",
              ")"
            ]
          },
          "metadata": {},
          "execution_count": 61
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "\n",
        "num_epochs = 10\n",
        "mse_multp = 0.5\n",
        "cls_multp = 0.5\n",
        "\n",
        "model.train()\n",
        "\n",
        "for epoch in range(num_epochs):\n",
        "     total_mseloss = 0.0\n",
        "     total_clsloss = 0.0\n",
        "     for ind, data in enumerate(train_loader ):\n",
        "         img, labels = data[0].to(device), data[1].to(device)\n",
        "         output, output_en = model(img)\n",
        "         loss_mse = distance(output, img)\n",
        "         loss_cls = class_loss(output_en, labels)\n",
        "         loss = (mse_multp * loss_mse) + (cls_multp * loss_cls)  # Combine two losses together\n",
        "         optimizer.zero_grad()\n",
        "         loss.backward()\n",
        "         optimizer.step()\n",
        "         # Track this epoch's loss\n",
        "         total_mseloss += loss_mse.item()\n",
        "         total_clsloss += loss_cls.item()\n",
        "\n",
        "     # Check accuracy on test set after each epoch:\n",
        "     model.eval()   # Turn off dropout in evaluation mode\n",
        "     acc = 0.0\n",
        "     total_samples = 0\n",
        "     for data in test_loader:\n",
        "         # We only care about the 10 dimensional encoder output for classification\n",
        "         img, labels = data[0].to(device), data[1].to(device)\n",
        "         _, output_en = model(img)\n",
        "         # output_en contains 10 values for each input, apply softmax to calculate class probabilities\n",
        "         prob = output_en\n",
        "         pred = torch.max(prob, dim=1)[1].detach().cpu().numpy() # Max prob assigned to class\n",
        "         acc += (pred == labels.cpu().numpy()).sum()\n",
        "         total_samples += labels.shape[0]\n",
        "     model.train()   # Enables dropout back again\n",
        "     print('epoch [{}/{}], loss_mse: {:.4f}  loss_cls: {:.4f}  Acc on test: {:.4f}'.format(epoch+1, num_epochs, total_mseloss / len(train_loader ), total_clsloss / len(train_loader ), acc / total_samples))"
      ],
      "metadata": {
        "id": "Fu4UnC_u5-x_",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "8126af2e-a298-4b53-c402-ade3bc6e24d3"
      },
      "execution_count": 62,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "<ipython-input-58-a2bddc038e92>:29: UserWarning: Implicit dimension choice for softmax has been deprecated. Change the call to include dim=X as an argument.\n",
            "  x1 = F.softmax(self.bottleneck(x))\n",
            "<ipython-input-58-a2bddc038e92>:30: UserWarning: Implicit dimension choice for softmax has been deprecated. Change the call to include dim=X as an argument.\n",
            "  x = F.softmax(self.bottleneck(x))\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "epoch [1/10], loss_mse: 0.7217  loss_cls: 1.5706  Acc on test: 0.9591\n",
            "epoch [2/10], loss_mse: 0.7083  loss_cls: 1.5019  Acc on test: 0.9647\n",
            "epoch [3/10], loss_mse: 0.7083  loss_cls: 1.4962  Acc on test: 0.9726\n",
            "epoch [4/10], loss_mse: 0.7082  loss_cls: 1.4931  Acc on test: 0.9764\n",
            "epoch [5/10], loss_mse: 0.7082  loss_cls: 1.4893  Acc on test: 0.9774\n",
            "epoch [6/10], loss_mse: 0.7081  loss_cls: 1.4860  Acc on test: 0.9773\n",
            "epoch [7/10], loss_mse: 0.7080  loss_cls: 1.4854  Acc on test: 0.9791\n",
            "epoch [8/10], loss_mse: 0.7080  loss_cls: 1.4827  Acc on test: 0.9814\n",
            "epoch [9/10], loss_mse: 0.7079  loss_cls: 1.4826  Acc on test: 0.9808\n",
            "epoch [10/10], loss_mse: 0.7079  loss_cls: 1.4815  Acc on test: 0.9805\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "import numpy as np\n",
        "import matplotlib.pyplot as plt\n",
        "\n",
        "model.eval()\n",
        "img, labels = list(train_loader)[0]\n",
        "img = img.to(device)\n",
        "output, output_en = model(img)\n",
        "inp = img[0:10, 0, :, :].squeeze().detach().cpu()\n",
        "out = output[0:10, 0, :, :].squeeze().detach().cpu()\n",
        "\n",
        "# Just some trick to concatenate first ten images next to each other\n",
        "inp = inp.permute(1,0,2).reshape(28, -1).numpy()\n",
        "out = out.permute(1,0,2).reshape(28, -1).numpy()\n",
        "combined = np.vstack([inp, out])\n",
        "\n",
        "plt.imshow(combined)\n",
        "plt.show()"
      ],
      "metadata": {
        "id": "CP7b2Nr8AZSD",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 255
        },
        "outputId": "47279594-9be2-403f-b19d-27e1d862698c"
      },
      "execution_count": 63,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "<ipython-input-58-a2bddc038e92>:29: UserWarning: Implicit dimension choice for softmax has been deprecated. Change the call to include dim=X as an argument.\n",
            "  x1 = F.softmax(self.bottleneck(x))\n",
            "<ipython-input-58-a2bddc038e92>:30: UserWarning: Implicit dimension choice for softmax has been deprecated. Change the call to include dim=X as an argument.\n",
            "  x = F.softmax(self.bottleneck(x))\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "To59p6pRKPlE"
      },
      "execution_count": null,
      "outputs": []
    }
  ]
}