dae_train.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import os\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "import copy\n",
    "import random\n",
    "import itertools\n",
    "import tqdm\n",
    "\n",
    "import time\n",
    "import torch\n",
    "import torch.nn as nn\n",
    "import torchvision\n",
    "from torch.utils.data import DataLoader\n",
    "from torchvision import transforms\n",
    "\n",
    "from IPython.display import display, clear_output"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Import the dataset from the .mat file\n",
    "import scipy.io\n",
    "\n",
    "dataset_mat = scipy.io.loadmat('MNIST.mat')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Convert the dataset in numpy\n",
    "digits = copy.deepcopy(dataset_mat['input_images'])\n",
    "labels = copy.deepcopy(dataset_mat['output_labels'].squeeze(1))\n",
    "\n",
    "for i, img in enumerate(digits):\n",
    "    temp = np.reshape(digits[i], (28, 28)).T\n",
    "    digits[i] = np.reshape(temp, (784))\n",
    "    \n",
    "dataset = list(zip(digits, labels))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x576 with 25 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "### Plot some sample\n",
    "plt.close('all')\n",
    "fig, axs = plt.subplots(5, 5, figsize=(8,8))\n",
    "for ax in axs.flatten():\n",
    "    img, lab = random.choice(dataset)\n",
    "    ax.imshow(img.reshape(28,28), cmap='gray_r')\n",
    "    ax.set_title('Label: %d' % lab)\n",
    "    ax.axis('off')\n",
    "plt.tight_layout()\n",
    "plt.savefig('digits.pdf', dpi = 300)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "def create_loader(dataset, batch_loader):\n",
    "    data, labels = zip(*dataset)\n",
    "\n",
    "    # Convert to torch\n",
    "    data = torch.tensor(data).float().view(-1, len(data), 28, 28).squeeze(0)\n",
    "    data = data[:, None, :, :]\n",
    "    labels = torch.tensor(labels).long().view(-1, len(labels)).squeeze(0)\n",
    "    dataset_torch = torch.utils.data.TensorDataset(data, labels.squeeze(0))\n",
    "\n",
    "    return DataLoader(dataset_torch, batch_size=batch_loader, shuffle = True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "def split_dataset(dataset, split):\n",
    "    random.shuffle(dataset)\n",
    "    datatrain = dataset[0:split[0]]\n",
    "    dataval = dataset[split[0]:split[1]]\n",
    "    datatest = dataset[split[1]:split[2]]\n",
    "    \n",
    "    return datatrain, dataval, datatest"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {},
   "outputs": [],
   "source": [
    "def add_noise(img, sigma, prob = 0.8):\n",
    "    rand_numbers = np.random.rand(len(img))\n",
    "    noised_img = copy.deepcopy(img)\n",
    "\n",
    "    for i in range(len(img)):\n",
    "        if rand_numbers[i] < prob:\n",
    "            noise = sigma * torch.randn(*img[i].shape)\n",
    "            noised_img[i] = np.clip(noised_img[i] + noise, 0., 1.)\n",
    "    return noised_img"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x576 with 25 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "### Plot some sample\n",
    "plt.close('all')\n",
    "fig, axs = plt.subplots(5, 5, figsize=(8,8))\n",
    "for ax in axs.flatten():\n",
    "    img, lab = random.choice(dataset)\n",
    "    noise = sigma * np.random.randn(784)\n",
    "    noised_img = np.clip(img + noise, 0., 1.)\n",
    "    \n",
    "    ax.imshow(noised_img.reshape(28,28), cmap='gray_r')\n",
    "    ax.set_title('Label: %d' % lab)\n",
    "    ax.axis('off')\n",
    "plt.tight_layout()\n",
    "plt.savefig('digits_noised.pdf', dpi = 300)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {},
   "outputs": [],
   "source": [
    "split = [40000, 50000, 60000]\n",
    "data = split_dataset(dataset, split)\n",
    "\n",
    "batch = [256, 10000, 10000]\n",
    "\n",
    "trainset = copy.deepcopy(data[0])\n",
    "valset = copy.deepcopy(data[1])\n",
    "testset = copy.deepcopy(data[2])\n",
    "\n",
    "trainloader = create_loader(trainset, batch[0])\n",
    "valloader = create_loader(valset, batch[1])\n",
    "testloader = create_loader(testset, batch[2])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "metadata": {},
   "outputs": [],
   "source": [
    "class Autoencoder(nn.Module):\n",
    "    \n",
    "    def __init__(self, encoded_space_dim):\n",
    "        super().__init__()\n",
    "        \n",
    "        ### Encoder\n",
    "        self.encoder_cnn = nn.Sequential(\n",
    "            nn.Conv2d(1, 8, 3, stride=2, padding=1),\n",
    "            nn.ReLU(True),\n",
    "            nn.Conv2d(8, 16, 3, stride=2, padding=1),\n",
    "            nn.ReLU(True),\n",
    "            nn.Conv2d(16, 32, 3, stride=2, padding=0),\n",
    "            nn.ReLU(True)\n",
    "        )\n",
    "        self.encoder_lin = nn.Sequential(\n",
    "            nn.Linear(3 * 3 * 32, 64),\n",
    "            nn.ReLU(True),\n",
    "            nn.Linear(64, encoded_space_dim)\n",
    "        )\n",
    "        \n",
    "        ### Decoder\n",
    "        self.decoder_lin = nn.Sequential(\n",
    "            nn.Linear(encoded_space_dim, 64),\n",
    "            nn.ReLU(True),\n",
    "            nn.Linear(64, 3 * 3 * 32),\n",
    "            nn.ReLU(True)\n",
    "        )\n",
    "        self.decoder_conv = nn.Sequential(\n",
    "            nn.ConvTranspose2d(32, 16, 3, stride=2, output_padding=0),\n",
    "            nn.ReLU(True),\n",
    "            nn.ConvTranspose2d(16, 8, 3, stride=2, padding=1, output_padding=1),\n",
    "            nn.ReLU(True),\n",
    "            nn.ConvTranspose2d(8, 1, 3, stride=2, padding=1, output_padding=1)\n",
    "        )\n",
    "\n",
    "    def forward(self, x):\n",
    "        x = self.encode(x)\n",
    "        x = self.decode(x)\n",
    "        return x\n",
    "\n",
    "    def encode(self, x):\n",
    "        # Apply convolutions\n",
    "        x = self.encoder_cnn(x)\n",
    "        # Flatten\n",
    "        x = x.view([x.size(0), -1])\n",
    "        # Apply linear layers\n",
    "        x = self.encoder_lin(x)\n",
    "        return x\n",
    "    \n",
    "    def decode(self, x):\n",
    "        # Apply linear layers\n",
    "        x = self.decoder_lin(x)\n",
    "        # Reshape\n",
    "        x = x.view([-1, 32, 3, 3])\n",
    "        # Apply transposed convolutions\n",
    "        x = self.decoder_conv(x)\n",
    "        x = torch.sigmoid(x)\n",
    "        return x"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {},
   "outputs": [],
   "source": [
    "def train_epoch(net, dataloader, loss_fn, optimizer, sigma, noising = True):\n",
    "    net.train()\n",
    "    avg_loss = 0\n",
    "    \n",
    "    for img_batch, lab_batch in dataloader:\n",
    "        if noising:\n",
    "            noised_img_batch = add_noise(img_batch, sigma)\n",
    "        noised_img_batch = noised_img_batch.to(device)\n",
    "        \n",
    "        output = net(noised_img_batch)\n",
    "                \n",
    "        # Loss between the output and the original images\n",
    "        loss = loss_fn(output, img_batch.to(device))\n",
    "        \n",
    "        optim.zero_grad()\n",
    "        loss.backward()\n",
    "        optim.step()\n",
    "        \n",
    "        avg_loss += loss.data\n",
    "        \n",
    "    return avg_loss/len(dataloader)\n",
    "\n",
    "def test_epoch(net, dataloader, loss_fn):\n",
    "    net.eval()\n",
    "    avg_loss = 0\n",
    "    \n",
    "    with torch.no_grad():\n",
    "        for img_batch, lab_batch in dataloader:\n",
    "            img_batch = img_batch.to(device)\n",
    "            output = net(img_batch)\n",
    "            \n",
    "            # The validation loss instead does not see noisy images\n",
    "            loss = loss_fn(output, img_batch)\n",
    "            avg_loss += loss.data            \n",
    "        \n",
    "    return avg_loss/len(dataloader)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {},
   "outputs": [],
   "source": [
    "def train_model(net, epochs, trainloader, valloader, loss_fn, optim, sigma):\n",
    "    train_loss_log = []\n",
    "    val_loss_log = []\n",
    "        \n",
    "    # Pick the best model\n",
    "    best_val_loss = 10000 \n",
    "    t0 = time.time()\n",
    "    for num_epoch in range(epochs):\n",
    "        time_start = time.time()\n",
    "        \n",
    "        train_loss = train_epoch(net, dataloader=trainloader, loss_fn=loss_fn, optimizer=optim, sigma = sigma)\n",
    "        train_loss_log.append(train_loss)\n",
    "        \n",
    "        val_loss = test_epoch(net, dataloader=valloader, loss_fn=loss_fn)\n",
    "        val_loss_log.append(val_loss)\n",
    "\n",
    "        time_end = time.time()\n",
    "        time_epoch = (time_end-time_start)/60\n",
    "        eta = time_epoch*(epochs-num_epoch)\n",
    "        \n",
    "        print('Epoch', num_epoch + 1)\n",
    "        print('[' + '#'*(100-int(100*(1 -num_epoch/epochs))) + ' '*int(100*(1 - num_epoch/epochs)) + ']')\n",
    "        print('\\t Training loss: ' + f'{train_loss:.5f}')\n",
    "        print('\\t Validation loss: ' + f'{val_loss:.5f}')\n",
    "        print('\\t %4.2f minutes left' % (eta))\n",
    "        print('\\t Elaps: %4.2f minutes' % ((time.time()-t0)/60))\n",
    "        clear_output(wait = True)\n",
    "        \n",
    "        if val_loss < best_val_loss:\n",
    "            best_model = net\n",
    "            best_val_loss = val_loss\n",
    "            best_train_loss = train_loss\n",
    "            epoch_stop = num_epoch\n",
    "            #print('Best model!')\n",
    "        #print('\\n')\n",
    "        \n",
    "    return best_model, train_loss_log, val_loss_log, epoch_stop"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {},
   "outputs": [],
   "source": [
    "def plot_rec(net, img, filename):\n",
    "    net.eval()\n",
    "    with torch.no_grad():\n",
    "        rec_img  = net(img)\n",
    "    fig, axs = plt.subplots(1, 2, figsize=(12,6))\n",
    "    \n",
    "    axs[0].imshow(img.cpu().squeeze().numpy(), cmap='gray_r')\n",
    "    axs[0].set_title('Original image')\n",
    "    axs[1].imshow(rec_img.cpu().squeeze().numpy(), cmap='gray_r')\n",
    "    axs[1].set_title('Reconstructed image')\n",
    "    axs[0].axis('off')\n",
    "    axs[1].axis('off')\n",
    "\n",
    "    plt.tight_layout()\n",
    "    plt.pause(0.1)\n",
    "    fig.savefig(filename, dpi = 300)\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Selected device: cuda\n"
     ]
    }
   ],
   "source": [
    "device = torch.device('cuda') if torch.cuda.is_available() else torch.device('cpu')\n",
    "print('Selected device:', device)\n",
    "\n",
    "num_epochs = 150"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x432 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "tvet_list = []\n",
    "encode_dim_list = [2,6,8,10,15,20, 30, 50, 70]\n",
    "\n",
    "\n",
    "sigma = 0.5\n",
    "\n",
    "for encode_dim in encode_dim_list:\n",
    "    ### Init\n",
    "    net = Autoencoder(encoded_space_dim=encode_dim)\n",
    "    loss_fn = torch.nn.MSELoss()\n",
    "    lr = 1e-3\n",
    "    optim = torch.optim.Adam(net.parameters(), lr=lr, weight_decay=1e-5)\n",
    "    \n",
    "    net.to(device)\n",
    "    \n",
    "    ### Training\n",
    "    net, tll, vll, estop = train_model(net, num_epochs, trainloader, valloader, loss_fn, optim, sigma)\n",
    "    test_loss = test_epoch(net, dataloader=testloader, loss_fn=loss_fn)\n",
    "    \n",
    "    ### Saving\n",
    "    torch.save(net.state_dict(), 'nets/conv_autoencoder_' + str(encode_dim) +'.torch')\n",
    "    tvet_list.append(tll)\n",
    "    tvet_list.append(vll)\n",
    "    tvet_list.append(estop)\n",
    "    tvet_list.append(test_loss)    \n",
    "    \n",
    "    ### Plots\n",
    "    plt.axhline(float(test_loss), color = 'red', label = 'Test loss', linestyle = '--')\n",
    "    plt.plot(tll, label = 'Training loss')\n",
    "    plt.plot(vll, label = 'Validation loss')\n",
    "    plt.axvline(x = estop, color ='g', linestyle='--', label = 'Best model')\n",
    "    plt.xlabel('Epoch')\n",
    "    plt.ylabel('Loss (encode dimension = ' + str(encode_dim) + ')')\n",
    "    \n",
    "    plt.legend()\n",
    "    plt.savefig('loss_' + str(encode_dim) + '.pdf', dpi = 300)\n",
    "    \n",
    "    idx = np.random.randint(len(testset))\n",
    "    img = torch.tensor(testset[idx][0].reshape(28,28))\n",
    "    img = img[None, None, :, :].to(device)\n",
    "    \n",
    "    plot_rec(net, img, 'rec_' + str(encode_dim) + '.pdf')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "metadata": {},
   "outputs": [],
   "source": [
    "np.save('tvet_list.npy', tvet_list, allow_pickle = True)\n",
    "np.save('testset', testset, allow_pickle = True)"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.9"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}