Untitled1.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "Untitled1.ipynb",
      "provenance": []
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "FSY-uw4MEghO"
      },
      "outputs": [],
      "source": [
        "import torch\n",
        "import argparse\n",
        "import numpy as np\n",
        "import torch.nn as nn\n",
        "import torch.optim as optim\n",
        "import torchvision.datasets as Data\n",
        "from torch.utils.data import DataLoader\n",
        "from torchvision import transforms\n",
        "from torch.autograd import Variable\n",
        "from torchvision.utils import save_image\n",
        "from torchvision.utils import make_grid\n",
        "\n",
        "transform = transforms.Compose([\n",
        "    transforms.ToTensor(),\n",
        "    transforms.Resize(size=(64, 64)),\n",
        "    transforms.Normalize([0.5], [0.5])\n",
        "])\n",
        "\n",
        "batch_size = 64\n",
        "\n",
        "data = Data.ImageFolder(root=\"/content/drive/MyDrive/New Folder\", transform=transform)\n",
        "load = DataLoader(data, batch_size, shuffle=True)\n",
        "\n",
        "\n",
        "def weights_init(m):\n",
        "    classname = m.__class__.__name__\n",
        "    if classname.find('Conv') != -1:\n",
        "        torch.nn.init.normal_(m.weight, 0.0, 0.02)\n",
        "    elif classname.find('BatchNorm') != -1:\n",
        "        torch.nn.init.normal_(m.weight, 1.0, 0.02)\n",
        "        torch.nn.init.zeros_(m.bias)\n",
        "\n",
        "\n",
        "class Discriminator(nn.Module):\n",
        "    def __init__(self, channels_img=3, channels_disc=64):\n",
        "        super(Discriminator, self).__init__()\n",
        "        self.network = nn.Sequential(\n",
        "            nn.Conv2d(in_channels=channels_img, out_channels=channels_disc, kernel_size=(4, 4), stride=(2, 2),\n",
        "                      padding=(1, 1)),\n",
        "            nn.BatchNorm2d(channels_disc),\n",
        "            nn.LeakyReLU(0.2),\n",
        "            nn.Conv2d(in_channels=channels_disc, out_channels=channels_disc * 2, kernel_size=(4, 4), padding=(1, 1),\n",
        "                      stride=(2, 2)),\n",
        "            nn.BatchNorm2d(channels_disc * 2),\n",
        "            nn.LeakyReLU(0.2),\n",
        "            nn.Conv2d(in_channels=channels_disc * 2, out_channels=channels_disc * 4, kernel_size=(4, 4), padding=(1, 1),\n",
        "                      stride=(2, 2)),\n",
        "            nn.BatchNorm2d(channels_disc * 4),\n",
        "            nn.LeakyReLU(0.2),\n",
        "            nn.Conv2d(in_channels=channels_disc * 4, out_channels=channels_disc * 8, kernel_size=(4, 4), padding=(1, 1),\n",
        "                      stride=(2, 2)),\n",
        "            nn.BatchNorm2d(channels_disc * 8),\n",
        "            nn.LeakyReLU(0.2),\n",
        "            nn.Conv2d(in_channels=channels_disc * 8, out_channels=1, kernel_size=(4, 4), padding=(0, 0), stride=(2, 2)),\n",
        "\n",
        "            nn.Sigmoid(),\n",
        "            nn.Flatten()\n",
        "        )\n",
        "\n",
        "    def forward(self, x):\n",
        "        x = self.network(x)\n",
        "        return x\n",
        "\n",
        "\n",
        "class Generator(nn.Module):\n",
        "    def __init__(self, z_dim=64, noise_dim=100, channels_img=3):\n",
        "        super(Generator, self).__init__()\n",
        "        self.network = nn.Sequential(\n",
        "            nn.ConvTranspose2d(in_channels=noise_dim, out_channels=z_dim * 16, kernel_size=(4, 4), stride=(1, 1),\n",
        "                               padding=(0, 0)),\n",
        "            nn.BatchNorm2d(z_dim * 16),\n",
        "            nn.ReLU(),\n",
        "            nn.ConvTranspose2d(in_channels=z_dim * 16, out_channels=z_dim * 8, kernel_size=(4, 4), stride=(2, 2),\n",
        "                               padding=(1, 1)),\n",
        "            nn.BatchNorm2d(z_dim * 8),\n",
        "            nn.ReLU(),\n",
        "            nn.ConvTranspose2d(in_channels=z_dim * 8, out_channels=z_dim * 4, kernel_size=(4, 4), stride=(2, 2),\n",
        "                               padding=(1, 1)),\n",
        "            nn.BatchNorm2d(z_dim * 4),\n",
        "            nn.ReLU(),\n",
        "            nn.ConvTranspose2d(in_channels=z_dim * 4, out_channels=z_dim * 2, kernel_size=(4, 4), stride=(2, 2),\n",
        "                               padding=(1, 1)),\n",
        "            nn.BatchNorm2d(z_dim * 2),\n",
        "            nn.ReLU(),\n",
        "            nn.ConvTranspose2d(in_channels=z_dim * 2, out_channels=channels_img, kernel_size=(4, 4), stride=(2, 2),\n",
        "                               padding=(1, 1)),\n",
        "            nn.Tanh()\n",
        "        )\n",
        "\n",
        "    def forward(self, x):\n",
        "        x = self.network(x)\n",
        "        # print(\"Gen: \", x.shape)\n",
        "        return x\n",
        "\n",
        "\n",
        "device = torch.device(\"cuda\" if torch.cuda.is_available() else \"cpu\")\n",
        "\n",
        "gen = Generator().to(device)\n",
        "gen.apply(weights_init)\n",
        "\n",
        "disc = Discriminator().to(device)\n",
        "disc.apply(weights_init)\n",
        "\n",
        "criterion = nn.BCELoss()\n",
        "\n",
        "\n",
        "learning_rate = 2e-4\n",
        "optimiser_disc = optim.Adam(disc.parameters(), learning_rate, betas=(0.5, 0.999))\n",
        "optimiser_gen = optim.Adam(gen.parameters(), learning_rate, betas=(0.5, 0.999))\n",
        "\n",
        "num_epochs = 20\n",
        "noise_dim = 100\n",
        "\n",
        "\n",
        "def generator_loss(fake_output, label):\n",
        "    gen_loss = criterion(fake_output, label)\n",
        "    return gen_loss\n",
        "\n",
        "\n",
        "def discriminator_loss(output, label):\n",
        "    disc_loss = criterion(output, label)\n",
        "    return disc_loss\n",
        "\n",
        "\n"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "from google.colab import drive\n",
        "drive.mount('/content/drive')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "9tzaBARrLYlj",
        "outputId": "422cbb50-c351-466f-ee0a-83d29756d0ff"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Drive already mounted at /content/drive; to attempt to forcibly remount, call drive.mount(\"/content/drive\", force_remount=True).\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "for epoch in range(20):\n",
        "    print(epoch)\n",
        "    for i, (real, _) in enumerate(load):\n",
        "        real = real.to(device)\n",
        "        noise = torch.randn((batch_size, noise_dim, 1, 1)).to(device)\n",
        "        fake = gen(noise)\n",
        "        if i >= 25 and i % 25 == 0:\n",
        "            print(i, len(load), 100 * i / len(load))\n",
        "\n",
        "        # TRAINING THE DISCRIMINATOR\n",
        "\n",
        "        real_target = Variable(torch.ones_like(disc(real)))\n",
        "        fake_target = Variable(torch.zeros_like(disc(fake)))\n",
        "\n",
        "        loss_disc_real = criterion(disc(real), real_target)\n",
        "        loss_disc_fake = criterion(disc(fake), fake_target)\n",
        "        loss_disc = (loss_disc_real + loss_disc_fake) / 2\n",
        "        optimiser_disc.zero_grad()\n",
        "        loss_disc.backward(retain_graph=True)\n",
        "        optimiser_disc.step()\n",
        "\n",
        "        # TRAINING THE GENERATOR\n",
        "\n",
        "        loss_gen = criterion(disc(fake), Variable(torch.ones_like(disc(fake))))\n",
        "        optimiser_gen.zero_grad()\n",
        "        loss_gen.backward()\n",
        "        optimiser_gen.step()\n",
        "\n",
        "    torch.save({\n",
        "    'epoch': epoch,\n",
        "    'model_state_dict': gen.state_dict()},\n",
        "    \"/content/drive/MyDrive/New Folder/Generator5.pth.tar\")\n",
        "\n",
        "    \n",
        "    random_img = torch.randn((batch_size, noise_dim, 1, 1)).to(device)\n",
        "    img = gen(random_img)\n",
        "    print(img.shape)\n",
        "    grid = make_grid(img)\n",
        "    print(grid.shape)\n",
        "    save_image(grid, \"output DCGAN gen5.jpg\")"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "KMfGXg5AJrGw",
        "outputId": "d6295e23-c56f-4a92-e6c9-35dab32317c9"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "0\n",
            "25 89 28.089887640449437\n",
            "50 89 56.17977528089887\n",
            "75 89 84.26966292134831\n",
            "torch.Size([64, 3, 64, 64])\n",
            "torch.Size([3, 530, 530])\n",
            "1\n",
            "25 89 28.089887640449437\n",
            "50 89 56.17977528089887\n",
            "75 89 84.26966292134831\n",
            "torch.Size([64, 3, 64, 64])\n",
            "torch.Size([3, 530, 530])\n",
            "2\n",
            "25 89 28.089887640449437\n",
            "50 89 56.17977528089887\n",
            "75 89 84.26966292134831\n",
            "torch.Size([64, 3, 64, 64])\n",
            "torch.Size([3, 530, 530])\n",
            "3\n",
            "25 89 28.089887640449437\n",
            "50 89 56.17977528089887\n",
            "75 89 84.26966292134831\n",
            "torch.Size([64, 3, 64, 64])\n",
            "torch.Size([3, 530, 530])\n",
            "4\n",
            "25 89 28.089887640449437\n",
            "50 89 56.17977528089887\n",
            "75 89 84.26966292134831\n",
            "torch.Size([64, 3, 64, 64])\n",
            "torch.Size([3, 530, 530])\n",
            "5\n",
            "25 89 28.089887640449437\n",
            "50 89 56.17977528089887\n",
            "75 89 84.26966292134831\n",
            "torch.Size([64, 3, 64, 64])\n",
            "torch.Size([3, 530, 530])\n",
            "6\n",
            "25 89 28.089887640449437\n",
            "50 89 56.17977528089887\n",
            "75 89 84.26966292134831\n",
            "torch.Size([64, 3, 64, 64])\n",
            "torch.Size([3, 530, 530])\n",
            "7\n",
            "25 89 28.089887640449437\n",
            "50 89 56.17977528089887\n",
            "75 89 84.26966292134831\n",
            "torch.Size([64, 3, 64, 64])\n",
            "torch.Size([3, 530, 530])\n",
            "8\n",
            "25 89 28.089887640449437\n",
            "50 89 56.17977528089887\n",
            "75 89 84.26966292134831\n",
            "torch.Size([64, 3, 64, 64])\n",
            "torch.Size([3, 530, 530])\n",
            "9\n",
            "25 89 28.089887640449437\n",
            "50 89 56.17977528089887\n",
            "75 89 84.26966292134831\n",
            "torch.Size([64, 3, 64, 64])\n",
            "torch.Size([3, 530, 530])\n",
            "10\n",
            "25 89 28.089887640449437\n",
            "50 89 56.17977528089887\n",
            "75 89 84.26966292134831\n",
            "torch.Size([64, 3, 64, 64])\n",
            "torch.Size([3, 530, 530])\n",
            "11\n",
            "25 89 28.089887640449437\n",
            "50 89 56.17977528089887\n",
            "75 89 84.26966292134831\n",
            "torch.Size([64, 3, 64, 64])\n",
            "torch.Size([3, 530, 530])\n",
            "12\n",
            "25 89 28.089887640449437\n",
            "50 89 56.17977528089887\n",
            "75 89 84.26966292134831\n",
            "torch.Size([64, 3, 64, 64])\n",
            "torch.Size([3, 530, 530])\n",
            "13\n",
            "25 89 28.089887640449437\n",
            "50 89 56.17977528089887\n",
            "75 89 84.26966292134831\n",
            "torch.Size([64, 3, 64, 64])\n",
            "torch.Size([3, 530, 530])\n",
            "14\n",
            "25 89 28.089887640449437\n",
            "50 89 56.17977528089887\n",
            "75 89 84.26966292134831\n",
            "torch.Size([64, 3, 64, 64])\n",
            "torch.Size([3, 530, 530])\n",
            "15\n",
            "25 89 28.089887640449437\n",
            "50 89 56.17977528089887\n",
            "75 89 84.26966292134831\n",
            "torch.Size([64, 3, 64, 64])\n",
            "torch.Size([3, 530, 530])\n",
            "16\n",
            "25 89 28.089887640449437\n",
            "50 89 56.17977528089887\n",
            "75 89 84.26966292134831\n",
            "torch.Size([64, 3, 64, 64])\n",
            "torch.Size([3, 530, 530])\n",
            "17\n",
            "25 89 28.089887640449437\n",
            "50 89 56.17977528089887\n",
            "75 89 84.26966292134831\n",
            "torch.Size([64, 3, 64, 64])\n",
            "torch.Size([3, 530, 530])\n",
            "18\n",
            "25 89 28.089887640449437\n",
            "50 89 56.17977528089887\n",
            "75 89 84.26966292134831\n",
            "torch.Size([64, 3, 64, 64])\n",
            "torch.Size([3, 530, 530])\n",
            "19\n",
            "25 89 28.089887640449437\n",
            "50 89 56.17977528089887\n",
            "75 89 84.26966292134831\n",
            "torch.Size([64, 3, 64, 64])\n",
            "torch.Size([3, 530, 530])\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "random_img = torch.randn((batch_size, noise_dim, 1, 1)).to(device)\n",
        "img = gen(random_img)\n",
        "print(img.shape)\n",
        "grid = make_grid(img)\n",
        "print(grid.shape)\n",
        "save_image(grid, \"output DCGAN 3.jpg\")"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "3UxDjKmydA-5",
        "outputId": "ffc5a263-2d9a-4961-ed03-a5b94cc0c4ba"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "torch.Size([64, 3, 64, 64])\n",
            "torch.Size([3, 530, 530])\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "for epoch in range(1, num_epochs + 1):\n",
        "    D_loss_list, G_loss_list = [], []\n",
        "    for index, (real_images, _) in enumerate(load):\n",
        "        real_images = real_images.to(device)\n",
        "        if (len(load)-index) % 30 == 0:\n",
        "          print(epoch, index, len(load))\n",
        "        optimiser_disc.zero_grad()\n",
        "        real_target = Variable(torch.ones(real_images.size(0))).to(device)\n",
        "        real_target = real_target.reshape(real_target.shape[0], 1)\n",
        "        fake_target = Variable(torch.zeros(real_images.size(0))).to(device)\n",
        "        fake_target = fake_target.reshape(fake_target.shape[0], 1)\n",
        "        output = disc(real_images)\n",
        "        # print(output.shape, real_target.shape[0])\n",
        "        D_real_loss = discriminator_loss(output, real_target)\n",
        "        D_real_loss.backward()\n",
        "        noise_vector = torch.randn(real_images.size(0), 100, 1, 1).to(device)\n",
        "        generated_image = gen(noise_vector)\n",
        "        output = disc(generated_image.detach())\n",
        "        D_fake_loss = discriminator_loss(output, fake_target)\n",
        "        # train with fake\n",
        "        D_fake_loss.backward()\n",
        "        D_total_loss = D_real_loss + D_fake_loss\n",
        "        D_loss_list.append(D_total_loss)\n",
        "        optimiser_disc.step()\n",
        "        optimiser_gen.zero_grad()\n",
        "        gen_output = disc(generated_image)\n",
        "        G_loss = generator_loss(gen_output, real_target)\n",
        "        G_loss_list.append(G_loss)\n",
        "        G_loss.backward()\n",
        "        optimiser_gen.step()\n",
        "\n",
        "    random_img = torch.randn((batch_size, noise_dim, 1, 1)).to(device)\n",
        "    img = gen(random_img)\n",
        "    print(img.shape)\n",
        "    grid = make_grid(img)\n",
        "    print(grid.shape)\n",
        "    save_image(grid, \"output DCGAN.jpg\")"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "iTqN0dqiVvHv",
        "outputId": "d1d7cc19-2a0e-4d85-b42a-169f6e8e0ac5"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "1 27 147\n",
            "1 57 147\n",
            "1 87 147\n",
            "1 117 147\n",
            "torch.Size([64, 3, 64, 64])\n",
            "torch.Size([3, 530, 530])\n",
            "2 27 147\n",
            "2 57 147\n",
            "2 87 147\n",
            "2 117 147\n",
            "torch.Size([64, 3, 64, 64])\n",
            "torch.Size([3, 530, 530])\n",
            "3 27 147\n",
            "3 57 147\n",
            "3 87 147\n",
            "3 117 147\n",
            "torch.Size([64, 3, 64, 64])\n",
            "torch.Size([3, 530, 530])\n",
            "4 27 147\n",
            "4 57 147\n",
            "4 87 147\n",
            "4 117 147\n",
            "torch.Size([64, 3, 64, 64])\n",
            "torch.Size([3, 530, 530])\n",
            "5 27 147\n",
            "5 57 147\n",
            "5 87 147\n",
            "5 117 147\n",
            "torch.Size([64, 3, 64, 64])\n",
            "torch.Size([3, 530, 530])\n",
            "6 27 147\n",
            "6 57 147\n",
            "6 87 147\n",
            "6 117 147\n",
            "torch.Size([64, 3, 64, 64])\n",
            "torch.Size([3, 530, 530])\n",
            "7 27 147\n",
            "7 57 147\n",
            "7 87 147\n",
            "7 117 147\n",
            "torch.Size([64, 3, 64, 64])\n",
            "torch.Size([3, 530, 530])\n",
            "8 27 147\n",
            "8 57 147\n",
            "8 87 147\n",
            "8 117 147\n",
            "torch.Size([64, 3, 64, 64])\n",
            "torch.Size([3, 530, 530])\n",
            "9 27 147\n",
            "9 57 147\n",
            "9 87 147\n",
            "9 117 147\n",
            "torch.Size([64, 3, 64, 64])\n",
            "torch.Size([3, 530, 530])\n",
            "10 27 147\n",
            "10 57 147\n",
            "10 87 147\n",
            "10 117 147\n",
            "torch.Size([64, 3, 64, 64])\n",
            "torch.Size([3, 530, 530])\n",
            "11 27 147\n",
            "11 57 147\n",
            "11 87 147\n",
            "11 117 147\n",
            "torch.Size([64, 3, 64, 64])\n",
            "torch.Size([3, 530, 530])\n",
            "12 27 147\n",
            "12 57 147\n",
            "12 87 147\n",
            "12 117 147\n",
            "torch.Size([64, 3, 64, 64])\n",
            "torch.Size([3, 530, 530])\n",
            "13 27 147\n",
            "13 57 147\n",
            "13 87 147\n",
            "13 117 147\n",
            "torch.Size([64, 3, 64, 64])\n",
            "torch.Size([3, 530, 530])\n",
            "14 27 147\n",
            "14 57 147\n",
            "14 87 147\n",
            "14 117 147\n",
            "torch.Size([64, 3, 64, 64])\n",
            "torch.Size([3, 530, 530])\n",
            "15 27 147\n",
            "15 57 147\n",
            "15 87 147\n",
            "15 117 147\n",
            "torch.Size([64, 3, 64, 64])\n",
            "torch.Size([3, 530, 530])\n",
            "16 27 147\n",
            "16 57 147\n",
            "16 87 147\n",
            "16 117 147\n",
            "torch.Size([64, 3, 64, 64])\n",
            "torch.Size([3, 530, 530])\n",
            "17 27 147\n",
            "17 57 147\n",
            "17 87 147\n",
            "17 117 147\n",
            "torch.Size([64, 3, 64, 64])\n",
            "torch.Size([3, 530, 530])\n",
            "18 27 147\n",
            "18 57 147\n",
            "18 87 147\n",
            "18 117 147\n",
            "torch.Size([64, 3, 64, 64])\n",
            "torch.Size([3, 530, 530])\n",
            "19 27 147\n",
            "19 57 147\n",
            "19 87 147\n",
            "19 117 147\n",
            "torch.Size([64, 3, 64, 64])\n",
            "torch.Size([3, 530, 530])\n",
            "20 27 147\n",
            "20 57 147\n",
            "20 87 147\n",
            "20 117 147\n",
            "torch.Size([64, 3, 64, 64])\n",
            "torch.Size([3, 530, 530])\n",
            "21 27 147\n",
            "21 57 147\n",
            "21 87 147\n",
            "21 117 147\n",
            "torch.Size([64, 3, 64, 64])\n",
            "torch.Size([3, 530, 530])\n",
            "22 27 147\n",
            "22 57 147\n",
            "22 87 147\n",
            "22 117 147\n",
            "torch.Size([64, 3, 64, 64])\n",
            "torch.Size([3, 530, 530])\n",
            "23 27 147\n",
            "23 57 147\n",
            "23 87 147\n",
            "23 117 147\n",
            "torch.Size([64, 3, 64, 64])\n",
            "torch.Size([3, 530, 530])\n",
            "24 27 147\n",
            "24 57 147\n",
            "24 87 147\n",
            "24 117 147\n",
            "torch.Size([64, 3, 64, 64])\n",
            "torch.Size([3, 530, 530])\n",
            "25 27 147\n",
            "25 57 147\n",
            "25 87 147\n",
            "25 117 147\n",
            "torch.Size([64, 3, 64, 64])\n",
            "torch.Size([3, 530, 530])\n",
            "26 27 147\n",
            "26 57 147\n",
            "26 87 147\n",
            "26 117 147\n",
            "torch.Size([64, 3, 64, 64])\n",
            "torch.Size([3, 530, 530])\n",
            "27 27 147\n",
            "27 57 147\n",
            "27 87 147\n",
            "27 117 147\n",
            "torch.Size([64, 3, 64, 64])\n",
            "torch.Size([3, 530, 530])\n",
            "28 27 147\n",
            "28 57 147\n",
            "28 87 147\n",
            "28 117 147\n",
            "torch.Size([64, 3, 64, 64])\n",
            "torch.Size([3, 530, 530])\n",
            "29 27 147\n",
            "29 57 147\n",
            "29 87 147\n",
            "29 117 147\n",
            "torch.Size([64, 3, 64, 64])\n",
            "torch.Size([3, 530, 530])\n",
            "30 27 147\n",
            "30 57 147\n",
            "30 87 147\n",
            "30 117 147\n",
            "torch.Size([64, 3, 64, 64])\n",
            "torch.Size([3, 530, 530])\n",
            "31 27 147\n",
            "31 57 147\n",
            "31 87 147\n",
            "31 117 147\n",
            "torch.Size([64, 3, 64, 64])\n",
            "torch.Size([3, 530, 530])\n",
            "32 27 147\n",
            "32 57 147\n",
            "32 87 147\n",
            "32 117 147\n",
            "torch.Size([64, 3, 64, 64])\n",
            "torch.Size([3, 530, 530])\n",
            "33 27 147\n",
            "33 57 147\n",
            "33 87 147\n",
            "33 117 147\n",
            "torch.Size([64, 3, 64, 64])\n",
            "torch.Size([3, 530, 530])\n",
            "34 27 147\n",
            "34 57 147\n",
            "34 87 147\n",
            "34 117 147\n",
            "torch.Size([64, 3, 64, 64])\n",
            "torch.Size([3, 530, 530])\n",
            "35 27 147\n",
            "35 57 147\n",
            "35 87 147\n",
            "35 117 147\n",
            "torch.Size([64, 3, 64, 64])\n",
            "torch.Size([3, 530, 530])\n",
            "36 27 147\n",
            "36 57 147\n",
            "36 87 147\n",
            "36 117 147\n",
            "torch.Size([64, 3, 64, 64])\n",
            "torch.Size([3, 530, 530])\n",
            "37 27 147\n",
            "37 57 147\n",
            "37 87 147\n",
            "37 117 147\n",
            "torch.Size([64, 3, 64, 64])\n",
            "torch.Size([3, 530, 530])\n",
            "38 27 147\n",
            "38 57 147\n",
            "38 87 147\n",
            "38 117 147\n",
            "torch.Size([64, 3, 64, 64])\n",
            "torch.Size([3, 530, 530])\n",
            "39 27 147\n",
            "39 57 147\n",
            "39 87 147\n",
            "39 117 147\n",
            "torch.Size([64, 3, 64, 64])\n",
            "torch.Size([3, 530, 530])\n",
            "40 27 147\n",
            "40 57 147\n",
            "40 87 147\n",
            "40 117 147\n",
            "torch.Size([64, 3, 64, 64])\n",
            "torch.Size([3, 530, 530])\n",
            "41 27 147\n",
            "41 57 147\n",
            "41 87 147\n",
            "41 117 147\n",
            "torch.Size([64, 3, 64, 64])\n",
            "torch.Size([3, 530, 530])\n",
            "42 27 147\n",
            "42 57 147\n",
            "42 87 147\n",
            "42 117 147\n",
            "torch.Size([64, 3, 64, 64])\n",
            "torch.Size([3, 530, 530])\n",
            "43 27 147\n",
            "43 57 147\n",
            "43 87 147\n",
            "43 117 147\n",
            "torch.Size([64, 3, 64, 64])\n",
            "torch.Size([3, 530, 530])\n",
            "44 27 147\n",
            "44 57 147\n",
            "44 87 147\n",
            "44 117 147\n",
            "torch.Size([64, 3, 64, 64])\n",
            "torch.Size([3, 530, 530])\n",
            "45 27 147\n",
            "45 57 147\n",
            "45 87 147\n",
            "45 117 147\n",
            "torch.Size([64, 3, 64, 64])\n",
            "torch.Size([3, 530, 530])\n",
            "46 27 147\n",
            "46 57 147\n",
            "46 87 147\n",
            "46 117 147\n",
            "torch.Size([64, 3, 64, 64])\n",
            "torch.Size([3, 530, 530])\n",
            "47 27 147\n",
            "47 57 147\n",
            "47 87 147\n",
            "47 117 147\n",
            "torch.Size([64, 3, 64, 64])\n",
            "torch.Size([3, 530, 530])\n",
            "48 27 147\n",
            "48 57 147\n",
            "48 87 147\n",
            "48 117 147\n",
            "torch.Size([64, 3, 64, 64])\n",
            "torch.Size([3, 530, 530])\n",
            "49 27 147\n",
            "49 57 147\n",
            "49 87 147\n",
            "49 117 147\n",
            "torch.Size([64, 3, 64, 64])\n",
            "torch.Size([3, 530, 530])\n",
            "50 27 147\n",
            "50 57 147\n",
            "50 87 147\n",
            "50 117 147\n",
            "torch.Size([64, 3, 64, 64])\n",
            "torch.Size([3, 530, 530])\n"
          ]
        }
      ]
    }
  ]
}