U47768_CH05_CNN(MNIST).ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "provenance": [],
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    },
    "accelerator": "GPU",
    "gpuClass": "standard"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/submouse9903/test-1/blob/master/U47768_CH05_CNN(MNIST).ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "pQm8v1zsccwY"
      },
      "outputs": [],
      "source": [
        "import torch \n",
        "import torch.nn as nn\n",
        "import torchvision\n",
        "import torchvision.transforms as transforms\n",
        "import matplotlib.pyplot as plt\n",
        "import numpy as np"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "train_dataset = torchvision.datasets.MNIST(root='./data', \n",
        "                                           train=True, \n",
        "                                           transform=transforms.ToTensor(),  \n",
        "                                           download=True)\n",
        "\n",
        "test_dataset = torchvision.datasets.MNIST(root='./data', \n",
        "                                          train=False, \n",
        "                                          transform=transforms.ToTensor())"
      ],
      "metadata": {
        "id": "UarI98xTcsjL",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "6b53a08e-a097-41ec-d6ab-745d6c305bec"
      },
      "execution_count": null,
      "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 ./data/MNIST/raw/train-images-idx3-ubyte.gz\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "100%|██████████| 9912422/9912422 [00:00<00:00, 343507235.58it/s]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Extracting ./data/MNIST/raw/train-images-idx3-ubyte.gz to ./data/MNIST/raw\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\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 ./data/MNIST/raw/train-labels-idx1-ubyte.gz\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "100%|██████████| 28881/28881 [00:00<00:00, 112370773.49it/s]\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Extracting ./data/MNIST/raw/train-labels-idx1-ubyte.gz to ./data/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 ./data/MNIST/raw/t10k-images-idx3-ubyte.gz\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "100%|██████████| 1648877/1648877 [00:00<00:00, 232224955.39it/s]\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Extracting ./data/MNIST/raw/t10k-images-idx3-ubyte.gz to ./data/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 ./data/MNIST/raw/t10k-labels-idx1-ubyte.gz\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "100%|██████████| 4542/4542 [00:00<00:00, 13030457.43it/s]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Extracting ./data/MNIST/raw/t10k-labels-idx1-ubyte.gz to ./data/MNIST/raw\n",
            "\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "i = 0\n",
        "image = train_dataset.data[i].float()/255\n",
        "label = train_dataset.targets[i]"
      ],
      "metadata": {
        "id": "QPa8wTZ0nYdH"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "plt.imshow(image, cmap='gray')\n",
        "plt.title('Label: %d' % label)\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 452
        },
        "id": "is-J2Dgznet6",
        "outputId": "1bcf0818-2e42-40ab-ede0-b3db508d5f18"
      },
      "execution_count": null,
      "outputs": [
        {
          "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": [
        "images = [train_dataset[i][0] for i in range(2)]\n",
        "# 타일 형식으로 이미지 시각화\n",
        "grid_img = torchvision.utils.make_grid(images, nrow=2, ncols = 1)\n",
        "plt.imshow(grid_img.permute(1, 2, 0))\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 318
        },
        "id": "sPc_jhYjoRj8",
        "outputId": "bf6f033e-9710-4ede-d9b4-ac5949172012"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "\n",
        "torchvision.utils.make_grid 함수가 흑백 이미지의 채널을 변화시켜서, (채널, 세로, 가로)이미지를 생성한다. 하지만 imshow 함수는 채널을 먼저 받을 수 없어서 (세로, 가로, 채널)로 데이터를 변화한다. permute 함수를 사용하여 텐서의 차원을 재배열함"
      ],
      "metadata": {
        "id": "76XC0GCcoyRf"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "print(\"dataset:\",train_dataset[i][0].shape)\n",
        "grid_img = torchvision.utils.make_grid(images, nrow=5, ncols = 5)\n",
        "print(\"grid.img:\", grid_img.shape)\n",
        "print(\"permute image\", grid_img.permute(1, 2, 0).shape)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "fZkJ1-uco5Y-",
        "outputId": "4bca5b9f-96f7-47c0-8e25-bd94670439be"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "dataset: torch.Size([1, 28, 28])\n",
            "grid.img: torch.Size([3, 32, 62])\n",
            "permute image torch.Size([32, 62, 3])\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "batchsize = 64\n",
        "train_loader = torch.utils.data.DataLoader(train_dataset, batchsize)"
      ],
      "metadata": {
        "id": "nVgrkjOIg_0U"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "Ryg8spZR00Vc"
      },
      "outputs": [],
      "source": [
        "x = train_dataset.data[0].float()/255\n",
        "x = x.reshape(1,28,28)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "g6KK9GB63YAy"
      },
      "outputs": [],
      "source": [
        "layer1 = nn.Sequential(nn.Conv2d(1, 32, kernel_size=3, stride=1, padding=1),\n",
        "                      nn.ReLU(),\n",
        "                      nn.MaxPool2d(kernel_size=2, stride=2))              "
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "h1 = nn.Conv2d(1, 32, kernel_size=3, stride=1, padding=1)(x)\n",
        "h1.shape"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "fSn0wqlND5WD",
        "outputId": "03ccfb82-27aa-43ed-e387-546c72e7dd21"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "torch.Size([32, 28, 28])"
            ]
          },
          "metadata": {},
          "execution_count": 6
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "h1 = nn.ReLU()(h1)\n",
        "h1.shape"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "keCD22iGEjef",
        "outputId": "e0fc7417-5f8b-48e6-b800-8decee6655c0"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "torch.Size([32, 28, 28])"
            ]
          },
          "metadata": {},
          "execution_count": 7
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "h1 = nn.MaxPool2d(kernel_size = 2, stride = 2)(h1)\n",
        "h1.shape"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "_xz2UT96Eqhq",
        "outputId": "45cef18c-e33e-4f9f-f376-8654bb52515d"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "torch.Size([32, 14, 14])"
            ]
          },
          "metadata": {},
          "execution_count": 8
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "layer1 = nn.Sequential(nn.Conv2d(1, 32, kernel_size=3, stride=1, padding=1),\n",
        "                      nn.ReLU(),\n",
        "                      nn.MaxPool2d(kernel_size=2, stride=2))              \n",
        "h1 = layer1(x)\n",
        "h1.shape"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "XPJvwU9aE9qy",
        "outputId": "55e09ddd-42d0-4c40-d199-fba916bd5e79"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "torch.Size([32, 14, 14])"
            ]
          },
          "metadata": {},
          "execution_count": 9
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "layer2 = nn.Sequential(nn.Conv2d(32, 64, kernel_size=3, stride=1, padding=1),\n",
        "                      nn.ReLU(),\n",
        "                      nn.MaxPool2d(kernel_size=2, stride=2))         \n",
        "h2 = layer2(h1)     \n",
        "h2.shape"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ijugYRpwFJvz",
        "outputId": "9891c283-f315-40ae-9cf0-d2352ad6db6d"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "torch.Size([64, 7, 7])"
            ]
          },
          "metadata": {},
          "execution_count": 10
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "layer3 = nn.Sequential(nn.Flatten(0,-1),\n",
        "                       nn.Linear(64*7*7,10),\n",
        "                       nn.Softmax(dim=0))\n",
        "output = layer3(h2)\n",
        "output.shape"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "lWb8KhDkFgrh",
        "outputId": "73e4bffe-5a26-42e2-f794-0c1d510856be"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "torch.Size([10])"
            ]
          },
          "metadata": {},
          "execution_count": 11
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "class MyNeuralNetwork(nn.Module):\n",
        "    def __init__(self):\n",
        "        super(MyNeuralNetwork, self).__init__()\n",
        "\n",
        "        self.layer1 = nn.Sequential(\n",
        "            nn.Conv2d(in_channels=1, out_channels=32, kernel_size=3, stride =1, padding= 1),\n",
        "            nn.ReLU(),\n",
        "            nn.MaxPool2d(kernel_size=2, stride=2)\n",
        "        )\n",
        "\n",
        "        self.layer2 = nn.Sequential(\n",
        "            nn.Conv2d(in_channels=32, out_channels=64, kernel_size=3, stride =1, padding= 1),\n",
        "            nn.ReLU(),\n",
        "            nn.MaxPool2d(kernel_size=2, stride=2)\n",
        "        )\n",
        "\n",
        "        self.layer3 = nn.Sequential(\n",
        "            nn.Linear(in_features=64*7*7, out_features=128, bias=True),\n",
        "            nn.ReLU(),\n",
        "            nn.Linear(in_features=128, out_features=64, bias=True),\n",
        "            nn.Sigmoid(),\n",
        "            nn.Linear(in_features=64, out_features=10, bias=True),\n",
        "            nn.Sigmoid(),\n",
        "            nn.Softmax(dim = 1)\n",
        "        )\n",
        "\n",
        "    def forward(self, x):\n",
        "        x = self.layer1(x)\n",
        "        x = self.layer2(x)\n",
        "        x = x.view(x.shape[0], -1)\n",
        "        x = self.layer3(x)\n",
        "        return x"
      ],
      "metadata": {
        "id": "3SeDtDrUJrRw"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "model = MyNeuralNetwork()\n",
        "for i, (data, target) in enumerate(train_loader):\n",
        "  output = model(data)\n",
        "  if i==0:\n",
        "    break\n",
        "output[0]"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "eoH40iUgKiyi",
        "outputId": "376f463c-53c9-4fb3-9cb5-4af51fff1690"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "tensor([0.0916, 0.1030, 0.1101, 0.1056, 0.0971, 0.0922, 0.1066, 0.0998, 0.0935,\n",
              "        0.1004], grad_fn=<SelectBackward0>)"
            ]
          },
          "metadata": {},
          "execution_count": 28
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "device = 'cuda' if torch.cuda.is_available() else 'cpu'\n",
        "\n",
        "torch.manual_seed(1)\n",
        "if device == 'cuda':\n",
        "    torch.cuda.manual_seed_all(1)"
      ],
      "metadata": {
        "id": "Dl-IUPkWyBC6"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "device"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        },
        "id": "-huyFBa9yCoC",
        "outputId": "924a90e1-4e89-4491-ef8b-ac45eb9e93ce"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "'cuda'"
            ],
            "application/vnd.google.colaboratory.intrinsic+json": {
              "type": "string"
            }
          },
          "metadata": {},
          "execution_count": 30
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "model = MyNeuralNetwork().to(device)\n"
      ],
      "metadata": {
        "id": "cHSGNpx7yUZ4"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "0tbFueiDHdvz"
      },
      "outputs": [],
      "source": [
        "learning_rate = 0.01\n",
        "optimizer = torch.optim.Adam(model.parameters(), lr=learning_rate)\n",
        "cross_entropy = nn.CrossEntropyLoss().to(device)  "
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ci7hNgG8HOvq",
        "outputId": "c9fce45e-6ee0-415a-cd42-b2d707ab8901"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Accuracy: 0.9388\n",
            "Accuracy: 0.9733\n",
            "Accuracy: 0.9695\n",
            "Accuracy: 0.9622\n",
            "Accuracy: 0.9539\n",
            "Learning rate is updated\n",
            "Accuracy: 0.9737\n",
            "Accuracy: 0.9748\n",
            "Accuracy: 0.9753\n",
            "Accuracy: 0.9748\n",
            "Accuracy: 0.9768\n"
          ]
        }
      ],
      "source": [
        "num_B = 10\n",
        "loss_vec = np.zeros(num_B)\n",
        "optimizer.zero_grad()\n",
        "\n",
        "for i in range(num_B):\n",
        "  val_vec = []\n",
        "  if i == 5:\n",
        "    print(\"Learning rate is updated\")\n",
        "    optimizer = torch.optim.Adam(model.parameters(), lr=learning_rate/2)\n",
        "  for j, (data, target) in enumerate(train_loader):\n",
        "    data = data.to(device)\n",
        "    target = target.to(device)\n",
        "    output = model(data)\n",
        "    loss = cross_entropy(output, target)\n",
        "    optimizer.zero_grad()\n",
        "    loss.backward()\n",
        "    optimizer.step()\n",
        "    predicted = torch.max(output.data,1)\n",
        "    v =((predicted.indices == target).sum()).item()/len(target)\n",
        "    val_vec.append(v)\n",
        "  print (\"Accuracy: {:.4f}\".format(np.array(val_vec).mean()))"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "test_loader = torch.utils.data.DataLoader(test_dataset, batchsize)\n",
        "for i, (dataset, target) in enumerate(test_loader):\n",
        "  print(dataset[0,:,:,:].shape)\n",
        "  break\n",
        " \n"
      ],
      "metadata": {
        "id": "GRSKeSqu6wZ4",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "17276e88-783a-4694-85d1-36fc2be6ce2b"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "torch.Size([1, 28, 28])\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "v = 0\n",
        "for i, (dataset, target) in enumerate(test_loader):\n",
        "  X = dataset.to(device)\n",
        "  Y = target.to(device)\n",
        "  output = model(X)\n",
        "  predicted = torch.max(output.data,1)\n",
        "  v = v + ((predicted.indices == Y).sum()).item()\n",
        "print(v)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "LiDn1h5xLjyJ",
        "outputId": "8b6a2bce-9114-48a8-bd55-27150dc5fd47"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "9796\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "\"Accuracy is {:.4f}%\".format(v/len(test_dataset.targets))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        },
        "id": "KstUog45633T",
        "outputId": "ec1e0f0e-6341-46cd-d436-1bad2faa8419"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "'Accuracy is 0.9796%'"
            ],
            "application/vnd.google.colaboratory.intrinsic+json": {
              "type": "string"
            }
          },
          "metadata": {},
          "execution_count": 45
        }
      ]
    }
  ]
}