DCGAN_Cahn_Hillard.ipynb

{
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/sedaoturak/dcgan_cahn-hilliard/blob/main/DCGAN_Cahn_Hillard.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Install dependencies"
      ],
      "metadata": {
        "id": "5gDR_JOdSbVG"
      }
    },
    {
      "cell_type": "code",
      "execution_count": 1,
      "metadata": {
        "id": "cqEBnmthpuuy"
      },
      "outputs": [],
      "source": [
        "import torch\n",
        "import torch.nn as nn\n",
        "import torch.nn.functional as F\n",
        "import torch.backends.cudnn as cudnn\n",
        "from torch.autograd.variable import Variable\n",
        "import torchvision\n",
        "from torchvision import datasets, models, transforms\n",
        "import matplotlib.pyplot as plt\n",
        "import numpy as np\n",
        "import random\n",
        "import tqdm.notebook"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Load data"
      ],
      "metadata": {
        "id": "FVQgHt-bSkh5"
      }
    },
    {
      "cell_type": "code",
      "execution_count": 2,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Yxy5_3FIp7ul",
        "outputId": "775e21ea-1c3c-4048-a3fd-6b78c906d96d"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Mounted at /content/drive\n"
          ]
        }
      ],
      "source": [
        "from google.colab import drive \n",
        "drive.mount('/content/drive')\n",
        "\n",
        "with open('/content/drive/MyDrive/toy_datasets/Mechanical_MNIST/Case1_input_patterns_64_x_64.txt', 'r') as f: \n",
        "  g = np.loadtxt(f)\n",
        "with open('/content/drive/MyDrive/toy_datasets/Mechanical_MNIST/Case1_strain_energy.txt', 'r') as f:\n",
        "  e = np.loadtxt(f)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 3,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 432
        },
        "id": "LEu85ax7rCAL",
        "outputId": "6496777c-0aad-4ec4-dde0-a3859a4a9fe4"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ],
      "source": [
        "# visualize an image from dataset\n",
        "img = g.reshape(37523,64,64)\n",
        "plt.imshow(img[9,:,:], interpolation='nearest', cmap='binary')\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 4,
      "metadata": {
        "id": "Mk8dKz_UrSBI",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "5c01f248-17cf-424b-a903-8a71152c6953"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "((37523, 64, 64), (37523,))"
            ]
          },
          "metadata": {},
          "execution_count": 4
        }
      ],
      "source": [
        "# normalize the strain energy values\n",
        "from sklearn.preprocessing import MinMaxScaler\n",
        "from sklearn.preprocessing import PowerTransformer\n",
        "\n",
        "pt = PowerTransformer()\n",
        "scaler = MinMaxScaler(feature_range=(0,1))\n",
        "\n",
        "e_scaled = scaler.fit_transform(e[:,6].reshape(-1, 1))\n",
        "e_power = pt.fit_transform(e[:,6].reshape(-1, 1))\n",
        "e_power_scaled = scaler.fit_transform(e_power)\n",
        "e_scaled.shape, e_scaled.ravel().shape\n",
        "\n",
        "images = g.reshape((37523,64,64))\n",
        "energies = e_power_scaled.ravel()\n",
        "\n",
        "images.shape, energies.shape"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "fig, axs = plt.subplots(figsize=(10,4))\n",
        "\n",
        "axs.hist(e[:,6], label=f'Energies (Unscaled)', alpha=0.5)\n",
        "axs.hist(e_scaled, label=f'Energies (Scaled)', alpha=0.5)\n",
        "# axs.hist(e_power, label=f'Energies (Power Trans.)', alpha=0.5)\n",
        "axs.hist(e_power_scaled, label=f'Energies (Power Trans.+Scaled)', alpha=0.5)\n",
        "axs.legend()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 385
        },
        "id": "7IPa_gNh0YeC",
        "outputId": "2f4edf5e-06f4-4fe5-863b-969fdfe70697"
      },
      "execution_count": 5,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.legend.Legend at 0x7f1baefb40d0>"
            ]
          },
          "metadata": {},
          "execution_count": 5
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 1000x400 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 5,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "GcNjRg3Oq67x",
        "outputId": "4f374a58-2032-4cdd-cfef-6322fc2bfc90"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "(37523, 64, 64)\n",
            "(37523, 64, 64, 1)\n",
            "(37523, 1, 64, 64)\n"
          ]
        }
      ],
      "source": [
        "from torch.utils.data import Dataset\n",
        "from torchvision import datasets\n",
        "from torchvision.transforms import ToTensor\n",
        "from torch.utils.data import random_split\n",
        "from torch.utils.data import DataLoader\n",
        "\n",
        "class CustomDataset(Dataset):\n",
        "  def __init__(self, imgs, labels, transform):\n",
        "    self.labels = labels\n",
        "    self.imgs = imgs\n",
        "    self.transform = transform\n",
        "\n",
        "  def __len__(self):\n",
        "    return len(self.imgs)\n",
        "\n",
        "  def __getitem__(self, index):\n",
        "    if self.labels is None:\n",
        "      x = self.imgs[index]\n",
        "      x = self.transform(x)\n",
        "      return x\n",
        "    else:\n",
        "      x = self.imgs[index]\n",
        "      x = self.transform(x)\n",
        "      y = self.labels[index]\n",
        "    return x, y\n",
        "\n",
        "b, w, h, = [*images.shape]\n",
        "print(images.shape)\n",
        "images_wb = np.repeat(images[..., np.newaxis], 1, -1)\n",
        "print(images_wb.shape)\n",
        "images_wb = images_wb.reshape(b, 1, w, h)\n",
        "print(images_wb.shape)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 6,
      "metadata": {
        "id": "zyK4mnBSq7YT"
      },
      "outputs": [],
      "source": [
        "transform = transforms.Compose([\n",
        "      # transforms.Resize((32, 32)),\n",
        "        ])\n",
        "\n",
        "# Datasets\n",
        "normalized_images = images_wb\n",
        "data = torch.tensor(normalized_images, dtype=torch.float32)\n",
        "target = torch.tensor(energies, dtype=torch.float32).reshape(-1,1)\n",
        "\n",
        "if __name__ == \"__main__\":\n",
        "  dataset = CustomDataset(data, target, transform)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 7,
      "metadata": {
        "id": "mvHrjW6grYTj"
      },
      "outputs": [],
      "source": [
        "# train-test split\n",
        "train_set_size = int(len(dataset)*0.8)\n",
        "val_set_size = len(dataset) - train_set_size\n",
        "\n",
        "trainset, valset  = random_split(dataset, [train_set_size, val_set_size])\n",
        "batch_size = 64\n",
        "trainloader = DataLoader(trainset, batch_size=batch_size, shuffle=True, num_workers= 2)\n",
        "valloader = DataLoader(valset, batch_size=batch_size, shuffle=True, num_workers= 2)"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Model"
      ],
      "metadata": {
        "id": "xQGlk5UOdUqH"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "https://github.com/Yuumna/GAN-PyTorch/blob/main/DCGAN_batch_normalization.ipynb"
      ],
      "metadata": {
        "id": "stCvEhv5Sv6_"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "def get_discriminator_block( input_channels, output_channels, kernel_size=4, stride=2, final_layer=False):\n",
        "  if not final_layer:\n",
        "    return nn.Sequential(\n",
        "      nn.Conv2d(input_channels, output_channels, kernel_size, stride),\n",
        "      nn.BatchNorm2d(output_channels),\n",
        "      nn.LeakyReLU(0.2, inplace=True) )\n",
        "  else: \n",
        "    return nn.Sequential(\n",
        "        nn.Conv2d(input_channels, output_channels, kernel_size, stride)\n",
        "    )"
      ],
      "metadata": {
        "id": "vfQPE-rcUoHe"
      },
      "execution_count": 71,
      "outputs": []
    },
    {
      "cell_type": "code",
      "execution_count": 72,
      "metadata": {
        "id": "3WnE4UARr8QO"
      },
      "outputs": [],
      "source": [
        "class Discriminator(nn.Module):\n",
        "  '''im_chan: the number of channels of the output image, a scalar\n",
        "              (MNIST is black-and-white, so 1 channel is the default)\n",
        "    hidden_dim: the inner dimension, a scalar\n",
        "  '''\n",
        "  def __init__(self, im_chan=1, hidden_dim=16):\n",
        "    super(Discriminator, self).__init__()\n",
        "    self.disc = nn.Sequential(\n",
        "        get_discriminator_block(im_chan, hidden_dim),\n",
        "        get_discriminator_block(hidden_dim, hidden_dim * 2),\n",
        "        get_discriminator_block(hidden_dim * 2, 1, final_layer=True) )\n",
        "\n",
        "  def forward(self, image):\n",
        "    disc_pred = self.disc(image)\n",
        "    return disc_pred.view(len(disc_pred), -1)"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "def get_generator_block(input_channels, output_channels, kernel_size=3, stride=2, final_layer=False):\n",
        "  if not final_layer:\n",
        "    return nn.Sequential(\n",
        "      nn.ConvTranspose2d(input_channels, output_channels, kernel_size=kernel_size, stride=stride),\n",
        "      nn.BatchNorm2d(output_channels),\n",
        "      nn.ReLU(inplace=True) )\n",
        "  else: \n",
        "    return nn.Sequential(\n",
        "      nn.ConvTranspose2d(input_channels, output_channels, kernel_size, stride),\n",
        "      nn.Sigmoid() )"
      ],
      "metadata": {
        "id": "xqnek3wFU_zd"
      },
      "execution_count": 73,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "class Generator(nn.Module):\n",
        "  '''z_dim: the dimension of the noise vector, a scalar\n",
        "     im_chan: the number of channels of the output image, a scalar\n",
        "              (MNIST is black-and-white, so 1 channel is the default)\n",
        "     hidden_dim: the inner dimension, a scalar\n",
        "  '''\n",
        "  def __init__(self, z_dim=10, im_chan=1, hidden_dim=64):\n",
        "    super(Generator, self).__init__()\n",
        "    self.z_dim = z_dim\n",
        "    self.gen = nn.Sequential(\n",
        "        get_generator_block(z_dim, hidden_dim * 4),\n",
        "        get_generator_block(hidden_dim * 4, hidden_dim * 2, kernel_size=4, stride=1),\n",
        "        get_generator_block(hidden_dim * 2, hidden_dim),\n",
        "        get_generator_block(hidden_dim, im_chan, kernel_size=4, final_layer=True) )\n",
        "\n",
        "  def forward(self, noise):\n",
        "    '''\n",
        "    Function for completing a forward pass of the generator: Given a noise tensor, \n",
        "    returns generated images.\n",
        "    Parameters:\n",
        "    noise: a noise tensor with dimensions (n_samples, z_dim)\n",
        "    '''\n",
        "    x = noise.view(len(noise), self.z_dim, 1, 1)\n",
        "    return self.gen(x)"
      ],
      "metadata": {
        "id": "Me51Dqaa_q5u"
      },
      "execution_count": 74,
      "outputs": []
    },
    {
      "cell_type": "code",
      "execution_count": 81,
      "metadata": {
        "id": "9O3o5mI5sEnW"
      },
      "outputs": [],
      "source": [
        "device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')\n",
        "\n",
        "z_dim = 64\n",
        "\n",
        "gen = Generator(z_dim).to(device)\n",
        "disc = Discriminator().to(device)\n",
        "\n",
        "criterion = torch.nn.MSELoss(reduction='mean')\n",
        "lr = 0.0002 #learning_rate\n",
        "beta_1 = 0.5 \n",
        "beta_2 = 0.999\n",
        "\n",
        "gen_opt = torch.optim.Adam(gen.parameters(), lr=lr, betas=(beta_1, beta_2))\n",
        "disc_opt = torch.optim.Adam(disc.parameters(), lr=lr, betas=(beta_1, beta_2))"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "def weights_init(m):\n",
        "    classname = m.__class__.__name__\n",
        "    if classname.find('Conv') != -1:\n",
        "        nn.init.normal_(m.weight.data, 0.0, 0.02)\n",
        "    elif classname.find('BatchNorm') != -1:\n",
        "        nn.init.normal_(m.weight.data, 1.0, 0.02)\n",
        "        nn.init.constant_(m.bias.data, 0)\n",
        "        \n",
        "gen = gen.apply(weights_init)\n",
        "disc = disc.apply(weights_init)"
      ],
      "metadata": {
        "id": "3ZKRZOlXVzLB"
      },
      "execution_count": 76,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "def get_noise(n_samples, z_dim, device='cpu'):\n",
        "  '''\n",
        "  Function for creating noise vectors: Given the dimensions (n_samples, z_dim),\n",
        "  creates a tensor of that shape filled with random numbers from the normal distribution.\n",
        "  Parameters:\n",
        "      n_samples: the number of samples to generate, a scalar\n",
        "      z_dim: the dimension of the noise vector, a scalar\n",
        "      device: the device type\n",
        "  '''\n",
        "  return torch.randn(n_samples,z_dim,device=device)"
      ],
      "metadata": {
        "id": "XiLctWVVWALg"
      },
      "execution_count": 77,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "def get_disc_loss(gen, disc, criterion, real, num_images, z_dim, device):\n",
        "  '''\n",
        "  Return the loss of the discriminator given inputs.\n",
        "  Parameters:\n",
        "      gen: the generator model, which returns an image given z-dimensional noise\n",
        "      disc: the discriminator model, which returns a single-dimensional prediction of real/fake\n",
        "      criterion: the loss function, which should be used to compare \n",
        "              the discriminator's predictions to the ground truth reality of the images \n",
        "              (e.g. fake = 0, real = 1)\n",
        "      real: a batch of real images\n",
        "      num_images: the number of images the generator should produce, \n",
        "              which is also the length of the real images\n",
        "      z_dim: the dimension of the noise vector, a scalar\n",
        "      device: the device type\n",
        "  Returns:\n",
        "      disc_loss: a torch scalar loss value for the current batch\n",
        "  '''\n",
        "  fake_noise = get_noise(num_images, z_dim, device=device)\n",
        "  fake = gen(fake_noise)\n",
        "  disc_fake_pred = disc(fake.detach())\n",
        "  disc_fake_loss = criterion(disc_fake_pred, torch.zeros_like(disc_fake_pred))\n",
        "  disc_real_pred = disc(real)\n",
        "  disc_real_loss = criterion(disc_real_pred, torch.ones_like(disc_real_pred))\n",
        "  disc_loss = (disc_fake_loss + disc_real_loss) / 2\n",
        "\n",
        "  return disc_loss"
      ],
      "metadata": {
        "id": "5sd7-qSlKBnn"
      },
      "execution_count": 78,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "def get_gen_loss(gen, disc, criterion, num_images, z_dim, device):\n",
        "  '''\n",
        "  Return the loss of the generator given inputs.\n",
        "  Parameters:\n",
        "      gen: the generator model, which returns an image given z-dimensional noise\n",
        "      disc: the discriminator model, which returns a single-dimensional prediction of real/fake\n",
        "      criterion: the loss function, which should be used to compare \n",
        "              the discriminator's predictions to the ground truth reality of the images \n",
        "              (e.g. fake = 0, real = 1)\n",
        "      num_images: the number of images the generator should produce, \n",
        "              which is also the length of the real images\n",
        "      z_dim: the dimension of the noise vector, a scalar\n",
        "      device: the device type\n",
        "  Returns:\n",
        "      gen_loss: a torch scalar loss value for the c1urrent batch\n",
        "  '''\n",
        "  fake_noise = get_noise(num_images, z_dim, device=device)\n",
        "  fake = gen(fake_noise)\n",
        "  disc_fake_pred = disc(fake)\n",
        "  gen_loss = criterion(disc_fake_pred, torch.ones_like(disc_fake_pred))\n",
        "\n",
        "  return gen_loss"
      ],
      "metadata": {
        "id": "f70Mq-CmWI4V"
      },
      "execution_count": 79,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "n_epochs = 50\n",
        "display_step = 500\n",
        "cur_step = 0\n",
        "mean_generator_loss = 0\n",
        "mean_discriminator_loss = 0\n",
        "\n",
        "# Keep track of losses\n",
        "disc_losses = []\n",
        "gen_losses = []\n",
        "\n",
        "test_generator = True # Whether the generator should be tested\n",
        "gen_loss = False\n",
        "error = False\n",
        "for epoch in range(n_epochs):\n",
        "  # Dataloader returns the batches\n",
        "  for real, _ in tqdm(trainloader):\n",
        "    cur_batch_size = len(real)\n",
        "\n",
        "    real = real.to(device)\n",
        "\n",
        "    ### Update discriminator ###\n",
        "    # Zero out the gradients before backpropagation\n",
        "    disc_opt.zero_grad()\n",
        "\n",
        "    # Calculate discriminator loss\n",
        "    disc_loss = get_disc_loss(gen, disc, criterion, real, cur_batch_size, z_dim, device)\n",
        "    disc_losses.append(disc_loss.item())\n",
        "\n",
        "    # Update gradients\n",
        "    disc_loss.backward(retain_graph=True)\n",
        "\n",
        "    # Update optimizer\n",
        "    disc_opt.step()\n",
        "\n",
        "    # For testing purposes, to keep track of the generator weights\n",
        "    if test_generator:\n",
        "      old_generator_weights = gen.gen[0][0].weight.detach().clone()\n",
        "\n",
        "    gen_opt.zero_grad()\n",
        "    gen_loss = get_gen_loss(gen, disc, criterion, cur_batch_size, z_dim, device)\n",
        "    gen_losses.append(gen_loss.item())\n",
        "    gen_loss.backward()\n",
        "    gen_opt.step()\n",
        "\n",
        "    # For testing purposes, to check that your code changes the generator weights\n",
        "    if test_generator:\n",
        "      try:\n",
        "        assert lr > 0.0000002 or (gen.gen[0][0].weight.grad.abs().max() < 0.0005 and epoch == 0)\n",
        "        assert torch.any(gen.gen[0][0].weight.detach().clone() != old_generator_weights)\n",
        "      except:\n",
        "        error = True\n",
        "        print(\"Runtime tests have failed\")\n",
        "\n",
        "    # Keep track of the average discriminator loss\n",
        "    mean_discriminator_loss += disc_loss.item() / display_step\n",
        "\n",
        "    # Keep track of the average generator loss\n",
        "    mean_generator_loss += gen_loss.item() / display_step\n",
        "\n",
        "    ### Visualization code ###\n",
        "    # if cur_step % display_step == 0 and cur_step > 0:\n",
        "    #   print(f\"Step {cur_step}: Generator loss: {mean_generator_loss}, discriminator loss: {mean_discriminator_loss}\")\n",
        "    #   fake_noise = get_noise(cur_batch_size, z_dim, device=device)\n",
        "    #   fake = gen(fake_noise)\n",
        "    #   show_tensor_images(fake)\n",
        "    #   show_tensor_images(real)\n",
        "    #   mean_generator_loss = 0\n",
        "    #   mean_discriminator_loss = 0\n",
        "    cur_step += 1"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "aNtDGyJuAcyj",
        "outputId": "57bc8dea-e947-4cbf-e5ca-9c378b6c0736"
      },
      "execution_count": 82,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "100%|██████████| 470/470 [00:06<00:00, 69.85it/s]\n",
            "100%|██████████| 470/470 [00:05<00:00, 85.54it/s]\n",
            "100%|██████████| 470/470 [00:06<00:00, 75.68it/s]\n",
            "100%|██████████| 470/470 [00:05<00:00, 82.78it/s]\n",
            "100%|██████████| 470/470 [00:05<00:00, 83.57it/s]\n",
            "100%|██████████| 470/470 [00:06<00:00, 75.46it/s]\n",
            "100%|██████████| 470/470 [00:05<00:00, 86.92it/s]\n",
            "100%|██████████| 470/470 [00:06<00:00, 72.53it/s]\n",
            "100%|██████████| 470/470 [00:06<00:00, 73.39it/s]\n",
            "100%|██████████| 470/470 [00:06<00:00, 71.90it/s]\n",
            "100%|██████████| 470/470 [00:05<00:00, 85.35it/s]\n",
            "100%|██████████| 470/470 [00:06<00:00, 75.21it/s]\n",
            "100%|██████████| 470/470 [00:05<00:00, 81.72it/s]\n",
            "100%|██████████| 470/470 [00:05<00:00, 85.38it/s]\n",
            "100%|██████████| 470/470 [00:06<00:00, 73.05it/s]\n",
            "100%|██████████| 470/470 [00:05<00:00, 85.53it/s]\n",
            "100%|██████████| 470/470 [00:06<00:00, 72.69it/s]\n",
            "100%|██████████| 470/470 [00:05<00:00, 85.31it/s]\n",
            "100%|██████████| 470/470 [00:06<00:00, 74.23it/s]\n",
            "100%|██████████| 470/470 [00:06<00:00, 70.40it/s]\n",
            "100%|██████████| 470/470 [00:06<00:00, 77.49it/s]\n",
            "100%|██████████| 470/470 [00:05<00:00, 81.21it/s]\n",
            "100%|██████████| 470/470 [00:05<00:00, 85.58it/s]\n",
            "100%|██████████| 470/470 [00:06<00:00, 75.55it/s]\n",
            "100%|██████████| 470/470 [00:05<00:00, 86.99it/s]\n",
            "100%|██████████| 470/470 [00:06<00:00, 73.40it/s]\n",
            "100%|██████████| 470/470 [00:05<00:00, 89.01it/s]\n",
            "100%|██████████| 470/470 [00:06<00:00, 75.92it/s]\n",
            "100%|██████████| 470/470 [00:05<00:00, 84.45it/s]\n",
            "100%|██████████| 470/470 [00:05<00:00, 86.87it/s]\n",
            "100%|██████████| 470/470 [00:07<00:00, 59.24it/s]\n",
            "100%|██████████| 470/470 [00:05<00:00, 86.83it/s]\n",
            "100%|██████████| 470/470 [00:06<00:00, 73.03it/s]\n",
            "100%|██████████| 470/470 [00:05<00:00, 87.73it/s]\n",
            "100%|██████████| 470/470 [00:06<00:00, 78.04it/s]\n",
            "100%|██████████| 470/470 [00:05<00:00, 80.72it/s]\n",
            "100%|██████████| 470/470 [00:05<00:00, 88.12it/s]\n",
            "100%|██████████| 470/470 [00:06<00:00, 74.06it/s]\n",
            "100%|██████████| 470/470 [00:05<00:00, 87.37it/s]\n",
            "100%|██████████| 470/470 [00:06<00:00, 74.13it/s]\n",
            "100%|██████████| 470/470 [00:05<00:00, 82.08it/s]\n",
            "100%|██████████| 470/470 [00:07<00:00, 67.01it/s]\n",
            "100%|██████████| 470/470 [00:05<00:00, 84.80it/s]\n",
            "100%|██████████| 470/470 [00:05<00:00, 84.48it/s]\n",
            "100%|██████████| 470/470 [00:06<00:00, 76.40it/s]\n",
            "100%|██████████| 470/470 [00:05<00:00, 90.20it/s]\n",
            "100%|██████████| 470/470 [00:06<00:00, 74.09it/s]\n",
            "100%|██████████| 470/470 [00:05<00:00, 85.08it/s]\n",
            "100%|██████████| 470/470 [00:06<00:00, 74.79it/s]\n",
            "100%|██████████| 470/470 [00:05<00:00, 84.26it/s]\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "fig, ax = plt.subplots(figsize=(9, 6))\n",
        "font_size=16\n",
        "tick_size=14\n",
        "\n",
        "ax.plot(disc_losses, '-', label='Discriminator (Train)', linewidth=3)\n",
        "ax.plot(gen_losses,'--', label='Generator (Validation)', linewidth=3)\n",
        "\n",
        "ax.set_ylabel('Loss', fontsize=font_size)\n",
        "ax.set_xlabel('Number of Epochs', fontsize=font_size)\n",
        "ax.set_yscale('log')\n",
        "\n",
        "ax.xaxis.set_tick_params(width=2, length=5, labelsize=tick_size, color='dimgrey')\n",
        "ax.yaxis.set_tick_params(width=2, length=5, labelsize=tick_size, color='dimgrey')\n",
        "ax.tick_params(which='both', width=2, length=6, top=True, right=True, direction='in', color='dimgrey')\n",
        "plt.setp(ax.spines.values(), linewidth=2, color='dimgrey')\n",
        "ax.legend(fontsize=font_size)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 565
        },
        "id": "ZItyZ2dSW9ra",
        "outputId": "8ae359a5-5478-480d-ebd4-059e20f18a9b"
      },
      "execution_count": 86,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.legend.Legend at 0x7f005de9ca60>"
            ]
          },
          "metadata": {},
          "execution_count": 86
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 900x600 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "p7PCGpJPspGr"
      },
      "outputs": [],
      "source": [
        "epochs = 1\n",
        "\n",
        "train_loss = []\n",
        "valid_loss = []\n",
        "for epoch in range(epochs):\n",
        "    print(f\"Epoch {epoch+1} of {epochs}\")\n",
        "    # beta annealing\n",
        "    if epoch<30:\n",
        "      beta=0\n",
        "    else:\n",
        "      beta=0.5e-4\n",
        "\n",
        "    ztrain, train_epoch_loss = train(trainloader, trainset, beta)\n",
        "    zval, valid_epoch_loss = validate(valloader, valset, beta)\n",
        "    train_loss.append(train_epoch_loss)\n",
        "    valid_loss.append(valid_epoch_loss)\n",
        "\n",
        "    print(f\"Train Loss: {train_epoch_loss:.4f}\")\n",
        "    print(f\"Val Loss: {valid_epoch_loss:.4f}\")"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# save model\n",
        "import os\n",
        "PATH = '/content/drive/MyDrive/Colab Notebooks'\n",
        "torch.save(model, os.path.join(PATH, f'mech_MNIST_entire.pt'))\n",
        "\n",
        "# model_scripted = torch.jit.script(model.cpu())\n",
        "# model_scripted.save(os.path.join(PATH, f'mech_MNIST_scripted.pt'))"
      ],
      "metadata": {
        "id": "xSTxdmOcX78g"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# load trained model\n",
        "import os\n",
        "PATH = '/content/drive/MyDrive/Colab Notebooks'\n",
        "model = torch.load(os.path.join(PATH, f'mech_MNIST_entire.pt'))\n",
        "device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')"
      ],
      "metadata": {
        "id": "Lzrkum9WYhYu"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Reconstructed Images"
      ],
      "metadata": {
        "id": "dAKC6naJUbVv"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "len(trainloader.dataset.indices)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "H1_DHNKCiof8",
        "outputId": "11dda86b-135a-4923-e98c-ed9422cc438b"
      },
      "execution_count": 87,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "30018"
            ]
          },
          "metadata": {},
          "execution_count": 87
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "original_images = []\n",
        "reconstructed_images = []\n",
        "z_train = np.zeros((0,16))\n",
        "idx_train = []\n",
        "strain_energy = np.zeros((0,1))\n",
        "for i, (d,t) in tqdm(enumerate(trainloader)):\n",
        "  d = d.to(device)\n",
        "  t = t.to(device)\n",
        "  z, _, _, out = generator(d,t)\n",
        "  original_images.append(d.cpu().detach().numpy())\n",
        "  reconstructed_images.append(out.cpu().detach().numpy())\n",
        "  z_train = np.vstack([z_train,z.cpu().detach().numpy()])\n",
        "  strain_energy = np.vstack([strain_energy, t.cpu().detach().numpy()])\n",
        "strain_energy = strain_energy.ravel()"
      ],
      "metadata": {
        "id": "z9xoDkpsb45q"
      },
      "execution_count": null,
      "outputs": []
    }
  ],
  "metadata": {
    "colab": {
      "provenance": [],
      "toc_visible": true,
      "authorship_tag": "ABX9TyMVrWyxqSq8YmxvS35xt7QJ",
      "include_colab_link": true
    },
    "gpuClass": "standard",
    "kernelspec": {
      "display_name": "Python 3",
      "name": "python3"
    },
    "language_info": {
      "name": "python"
    },
    "accelerator": "GPU"
  },
  "nbformat": 4,
  "nbformat_minor": 0
}