adversarial-training.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "o7_vy8HV_NIr",
   "metadata": {
    "id": "o7_vy8HV_NIr"
   },
   "source": [
    "#Imports"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "81dc370f",
   "metadata": {
    "id": "81dc370f"
   },
   "outputs": [],
   "source": [
    "import torch\n",
    "import torchvision\n",
    "import torchvision.transforms as transforms\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "import torch.nn as nn\n",
    "import torch.nn.functional as F\n",
    "import torch.optim as optim\n",
    "from skimage.util import random_noise\n",
    "import random"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "IgdPSTDX_YVI",
   "metadata": {
    "id": "IgdPSTDX_YVI"
   },
   "source": [
    "#Preparer le GPU"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "1addda1b",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "1addda1b",
    "outputId": "cb946358-a353-41b0-d6c2-3e5eb3328bd5"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "NVIDIA GeForce RTX 3080 Ti\n"
     ]
    }
   ],
   "source": [
    "print(torch.cuda.get_device_name(0))\n",
    "device = torch.device(\"cuda\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "1XlKd7Yz_esY",
   "metadata": {
    "id": "1XlKd7Yz_esY"
   },
   "source": [
    "#Preparer le Dataset"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6vuQlxU-Jjxs",
   "metadata": {
    "id": "6vuQlxU-Jjxs"
   },
   "source": [
    "On recupère le dataset CIFAR10 et on le divise en deux sous-datasets:\n",
    "- L'ensemble d'apprentissage, commposé de 50 000 images\n",
    "- L'ensemble de test, composé de 10 000 images\n",
    "\n",
    "On va normaliser les images afin de normaliser les valeurs des pixels dans l'intervalle [0,1]. Pour cela nous avons precisé les moyennes et les ecarts types du dataset.\n",
    "\n",
    "Enfin nous avons utilisé un loader par sous-dataset en precisant la batch size de chacun.\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "47feb64e",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 101,
     "referenced_widgets": [
      "0791008ba84b432492c984cb0cec8422",
      "70752c805d8f45e0a52db2636a321319",
      "79a343f676314d5ba3a5b2df9422595a",
      "0dda7b18e0204c71927d29e3996553ae",
      "5d6a358b29d646c2a1e2bf80b4a7b9fc",
      "daa90939e369449c92587365fd0c3490",
      "9a03fd5d95614d86b1e501282c04c679",
      "f0e8828a28294f3d9c444b7c948e8f3d",
      "182d6ff5143543218e36f5e4e28de773",
      "bb1791ae91fc48eca9d628a26daf7f84",
      "0892224ca7b94e42a77b4427d40f354e"
     ]
    },
    "id": "47feb64e",
    "outputId": "4a11144f-6c36-42d5-c98a-802f3c44f9d5"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Downloading https://www.cs.toronto.edu/~kriz/cifar-10-python.tar.gz to ./data/cifar-10-python.tar.gz\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "0791008ba84b432492c984cb0cec8422",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/170498071 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Extracting ./data/cifar-10-python.tar.gz to ./data\n",
      "Files already downloaded and verified\n"
     ]
    }
   ],
   "source": [
    "transform = transforms.Compose([\n",
    "    transforms.ToTensor(),\n",
    "    transforms.Normalize((0.4914, 0.4822, 0.4465), (0.2023, 0.1994, 0.2010)),\n",
    "])\n",
    "\n",
    "batch_size = 64\n",
    "\n",
    "trainset = torchvision.datasets.CIFAR10(\n",
    "    root='./data', train=True, download=True, transform=transform)\n",
    "train_loader = torch.utils.data.DataLoader(\n",
    "    trainset, batch_size=batch_size, shuffle=True, num_workers=2)\n",
    "\n",
    "testset = torchvision.datasets.CIFAR10(\n",
    "    root='./data', train=False, download=True, transform=transform)\n",
    "test_loader = torch.utils.data.DataLoader(\n",
    "    testset, batch_size=100, shuffle=False, num_workers=2)\n",
    "\n",
    "classes = ('plane', 'car', 'bird', 'cat', 'deer',\n",
    "           'dog', 'frog', 'horse', 'ship', 'truck')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "d4fe5558",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 156
    },
    "id": "d4fe5558",
    "outputId": "107fcbf3-8c18-4cd9-a7b1-7263699e3853"
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "frog      deer      truck     plane     horse\n"
     ]
    }
   ],
   "source": [
    "#Functions to show an image\n",
    "def imshow(img):\n",
    "    img = img / 2 + 0.5     # unnormalize\n",
    "    npimg = img.numpy()\n",
    "    plt.imshow(np.transpose(npimg, (1, 2, 0)))\n",
    "    plt.show()\n",
    "\n",
    "#Get some random training images\n",
    "num_img = 5\n",
    "example_loader = torch.utils.data.DataLoader(trainset, batch_size=num_img, shuffle=True, num_workers=2)\n",
    "dataiter = iter(example_loader)\n",
    "images, labels = dataiter.next()\n",
    "\n",
    "#Show images\n",
    "imshow(torchvision.utils.make_grid(images))\n",
    "# print labels\n",
    "print('     '.join(f'{classes[labels[j]]:5s}' for j in range(num_img)))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "yGh92aFCArud",
   "metadata": {
    "id": "yGh92aFCArud"
   },
   "source": [
    "#Architecture du modèle"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "-1XTh4V4BJWv",
   "metadata": {
    "id": "-1XTh4V4BJWv"
   },
   "source": [
    "Pour notre modèle nous avons opté pour l'achitecture VGG. Nous avons prévu plusieurs configurations (VGG11, VG13, VGG16 et VGG19). \n",
    "\n",
    "L'architecture du modèle VGG alterne des couches de convolution et des couches de maxpooling."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "86c7b97b",
   "metadata": {
    "id": "86c7b97b",
    "outputId": "95d9b8dd-a43e-4a6f-e655-3fabb42ac2db"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "torch.Size([2, 10])\n",
      "VGG(\n",
      "  (features): Sequential(\n",
      "    (0): Conv2d(3, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "    (1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "    (2): ReLU(inplace=True)\n",
      "    (3): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "    (4): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "    (5): ReLU(inplace=True)\n",
      "    (6): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)\n",
      "    (7): Conv2d(64, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "    (8): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "    (9): ReLU(inplace=True)\n",
      "    (10): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "    (11): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "    (12): ReLU(inplace=True)\n",
      "    (13): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)\n",
      "    (14): Conv2d(128, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "    (15): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "    (16): ReLU(inplace=True)\n",
      "    (17): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "    (18): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "    (19): ReLU(inplace=True)\n",
      "    (20): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "    (21): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "    (22): ReLU(inplace=True)\n",
      "    (23): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)\n",
      "    (24): Conv2d(256, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "    (25): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "    (26): ReLU(inplace=True)\n",
      "    (27): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "    (28): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "    (29): ReLU(inplace=True)\n",
      "    (30): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "    (31): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "    (32): ReLU(inplace=True)\n",
      "    (33): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)\n",
      "    (34): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "    (35): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "    (36): ReLU(inplace=True)\n",
      "    (37): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "    (38): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "    (39): ReLU(inplace=True)\n",
      "    (40): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "    (41): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "    (42): ReLU(inplace=True)\n",
      "    (43): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)\n",
      "    (44): AvgPool2d(kernel_size=1, stride=1, padding=0)\n",
      "  )\n",
      "  (classifier): Sequential(\n",
      "    (0): Dropout(p=0.5, inplace=False)\n",
      "    (1): Linear(in_features=512, out_features=512, bias=True)\n",
      "    (2): ReLU(inplace=True)\n",
      "    (3): Dropout(p=0.5, inplace=False)\n",
      "    (4): Linear(in_features=512, out_features=512, bias=True)\n",
      "    (5): ReLU(inplace=True)\n",
      "    (6): Linear(in_features=512, out_features=10, bias=True)\n",
      "  )\n",
      ")\n"
     ]
    }
   ],
   "source": [
    "#Different architectures of vgg. Represent all the layers\n",
    "cfg = {\n",
    "    'VGG11': [64, 'M', 128, 'M', 256, 256, 'M', 512, 512, 'M', 512, 512, 'M'],\n",
    "    'VGG13': [64, 64, 'M', 128, 128, 'M', 256, 256, 'M', 512, 512, 'M', 512, 512, 'M'],\n",
    "    'VGG16': [64, 64, 'M', 128, 128, 'M', 256, 256, 256, 'M', 512, 512, 512, 'M', 512, 512, 512, 'M'],\n",
    "    'VGG19': [64, 64, 'M', 128, 128, 'M', 256, 256, 256, 256, 'M', 512, 512, 512, 512, 'M', 512, 512, 512, 512, 'M'],\n",
    "}\n",
    "\n",
    "\n",
    "#Model class\n",
    "class VGG(nn.Module):\n",
    "    def __init__(self, vgg_name):\n",
    "        super(VGG, self).__init__()\n",
    "        self.features = self._make_layers(cfg[vgg_name])\n",
    "        self.classifier = nn.Sequential(\n",
    "            nn.Dropout(),\n",
    "            nn.Linear(512, 512),\n",
    "            nn.ReLU(True),\n",
    "            nn.Dropout(),\n",
    "            nn.Linear(512, 512),\n",
    "            nn.ReLU(True),\n",
    "            nn.Linear(512, 10),\n",
    "        )\n",
    "\n",
    "    def forward(self, x):\n",
    "        out = self.features(x)\n",
    "        out = out.view(out.size(0), -1)\n",
    "        out = self.classifier(out)\n",
    "        return out\n",
    "\n",
    "    #For each element in the config selectionned, we create the layer associated \n",
    "    def _make_layers(self, cfg):\n",
    "        layers = []\n",
    "        in_channels = 3\n",
    "        for x in cfg:\n",
    "            if x == 'M':\n",
    "                layers += [nn.MaxPool2d(kernel_size=2, stride=2)]\n",
    "            else:\n",
    "                layers += [nn.Conv2d(in_channels, x, kernel_size=3, padding=1, bias=False),\n",
    "                           nn.BatchNorm2d(x),\n",
    "                           nn.ReLU(inplace=True)]\n",
    "                in_channels = x\n",
    "        layers += [nn.AvgPool2d(kernel_size=1, stride=1)]\n",
    "        return nn.Sequential(*layers)\n",
    "\n",
    "\n",
    "def test():\n",
    "    net = VGG('VGG16')\n",
    "    x = torch.randn(2,3,32,32)\n",
    "    y = net(x)\n",
    "    print(y.size())\n",
    "    print(net)\n",
    "\n",
    "test()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a51qOgBwQmis",
   "metadata": {
    "id": "a51qOgBwQmis"
   },
   "source": [
    "#Entrainement du modèle"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "860634d4",
   "metadata": {
    "id": "860634d4"
   },
   "outputs": [],
   "source": [
    "def training(net, criterion, optimizer, scheduler, n_epoch, attack=None, noise=None):\n",
    "    net.to(device)\n",
    "    for epoch in range(n_epoch):\n",
    "\n",
    "        running_loss = 0.0\n",
    "        test_loss = 0.0\n",
    "\n",
    "        ##TRAIN\n",
    "        for i, data in enumerate(train_loader, 0):\n",
    "            #Get the inputs; data is a list of [images, labels]\n",
    "            images, labels = data\n",
    "            #Put images and labels on GPU\n",
    "            images = images.to(device)\n",
    "            labels = labels.to(device)\n",
    "            \n",
    "            #Attack the images if precised (Adversarial Training)\n",
    "            if attack != None:\n",
    "                images = attack.compute(images, labels)\n",
    "\n",
    "            #Add noise if precised\n",
    "            if noise != None:\n",
    "                images = add_noise(images)\n",
    "    \n",
    "\n",
    "            #Zero the parameter gradients\n",
    "            optimizer.zero_grad(set_to_none=True)\n",
    "            #Forward + backward + optimize\n",
    "            outputs = net(images)\n",
    "            loss = criterion(outputs, labels)\n",
    "            loss.backward()\n",
    "            optimizer.step()\n",
    "            #Compute loss\n",
    "            running_loss += loss.item()*images.size(0)\n",
    "            \n",
    "        ''' IT ADD TOO MUCH TIME TO THE TRAINING\n",
    "        ##TEST\n",
    "        with torch.no_grad():\n",
    "            for i, data in enumerate(test_loader, 0):\n",
    "                # get the inputs; data is a list of [inputs, labels]\n",
    "                images, labels = data\n",
    "                images = images.to(device)\n",
    "                labels = labels.to(device)\n",
    "                # zero the parameter gradients\n",
    "                optimizer.zero_grad()\n",
    "                # forward + backward + optimize\n",
    "                outputs = net(images)\n",
    "                loss = criterion(outputs, labels)\n",
    "                # print statistics\n",
    "                test_loss += loss.item()*images.size(0)'''\n",
    "\n",
    "        scheduler.step()\n",
    "        print(f'epoch {epoch + 1} -> training_loss: {running_loss/50000:.3f}')#, test_loss: {test_loss/10000:.3f}')\n",
    "    print('Finished Training')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "a5426d04",
   "metadata": {
    "id": "a5426d04"
   },
   "outputs": [],
   "source": [
    "def save_model(model, path):\n",
    "    torch.save(model.state_dict(), path)\n",
    "\n",
    "def load_model(model, path):\n",
    "    model.load_state_dict(torch.load(path))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "334c22d6",
   "metadata": {
    "id": "334c22d6",
    "outputId": "7921e5c0-7b38-4786-962f-3a132683cbba"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training without attack :\n",
      "epoch 1 -> training_loss: 1.568\n",
      "epoch 2 -> training_loss: 1.014\n",
      "epoch 3 -> training_loss: 0.772\n",
      "epoch 4 -> training_loss: 0.608\n",
      "epoch 5 -> training_loss: 0.481\n",
      "epoch 6 -> training_loss: 0.383\n",
      "epoch 7 -> training_loss: 0.303\n",
      "epoch 8 -> training_loss: 0.245\n",
      "epoch 9 -> training_loss: 0.198\n",
      "epoch 10 -> training_loss: 0.156\n",
      "epoch 11 -> training_loss: 0.129\n",
      "epoch 12 -> training_loss: 0.108\n",
      "epoch 13 -> training_loss: 0.092\n",
      "epoch 14 -> training_loss: 0.076\n",
      "epoch 15 -> training_loss: 0.071\n",
      "epoch 16 -> training_loss: 0.057\n",
      "epoch 17 -> training_loss: 0.050\n",
      "epoch 18 -> training_loss: 0.051\n",
      "epoch 19 -> training_loss: 0.047\n",
      "epoch 20 -> training_loss: 0.038\n",
      "Finished Training\n",
      "Wall time: 8min 26s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "model = VGG(\"VGG16\")\n",
    "#Put the model on GPU\n",
    "model = model.to(device)\n",
    "\n",
    "criterion = nn.CrossEntropyLoss()\n",
    "optimizer = optim.SGD(model.parameters(), lr=0.001, momentum=0.9)\n",
    "scheduler = torch.optim.lr_scheduler.CosineAnnealingLR(optimizer, T_max=200)\n",
    "\n",
    "n_epoch = 20\n",
    "\n",
    "print(\"Training without attack :\")\n",
    "training(model, criterion, optimizer, scheduler, n_epoch)\n",
    "save_model(model, \"./vgg16.pth\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "QiE63EswRogw",
   "metadata": {
    "id": "QiE63EswRogw"
   },
   "source": [
    "#Evaluation du modèle"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "cd02ff77",
   "metadata": {
    "id": "cd02ff77"
   },
   "outputs": [],
   "source": [
    "def eval_model(model, loader, attack=None, noise=None):    \n",
    "    total =0\n",
    "    correct =0\n",
    "\n",
    "    for data in test_loader:\n",
    "        images, labels = data\n",
    "        images = images.to(device)\n",
    "        labels = labels.to(device)\n",
    "        if attack != None:\n",
    "            images = attack.compute(images, labels)\n",
    "        if noise != None:\n",
    "            images = add_noise(images)\n",
    "\n",
    "\n",
    "        outputs = model(images)\n",
    "        _, predicted = torch.max(outputs.data, 1)\n",
    "        total += labels.size(0)\n",
    "        correct += (predicted == labels).sum().item()\n",
    "\n",
    "    \n",
    "    print(\"Correct :\",correct)\n",
    "    print(\"Total :\", total)\n",
    "    print(f'Accuracy: {100 * correct // total} %')\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "e1ed3ff4",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "e1ed3ff4",
    "outputId": "f63deb56-766c-47d3-ca1f-bf534ae4e191"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "__________________________\n",
      "Eval on clean images\n",
      "8231\n",
      "10000\n",
      "Accuracy: 82 %\n"
     ]
    }
   ],
   "source": [
    "model = VGG(\"VGG16\")\n",
    "model.load_state_dict(torch.load('./vgg16.pth'))\n",
    "model = model.to(device)\n",
    "      \n",
    "print(\"Eval on clean images :\")\n",
    "eval_model(model, test_loader)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ke4-RpI1SV1X",
   "metadata": {
    "id": "ke4-RpI1SV1X"
   },
   "source": [
    "On voit que les resutlats de notre modèles sont très bon !\n",
    "\n",
    "Mais que se passerait-il si notre ensemble d'images de test venait à être alteré avec une attaque PGD ? Le modèle va-t-il être robuste ?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "64ee5899",
   "metadata": {
    "id": "64ee5899"
   },
   "outputs": [],
   "source": [
    "class ProjectedGradientDescent:\n",
    "    \"\"\"\n",
    "    A class used to attack the images\n",
    "\n",
    "    Methods\n",
    "    -------\n",
    "    compute(images, labels)\n",
    "        Compute all the adversarial images\n",
    "    \"\"\"\n",
    "  \n",
    "    def __init__(self, model, eps, alpha, num_iter):\n",
    "        self.model = model\n",
    "        self.eps = eps\n",
    "        self.alpha=alpha\n",
    "        self.num_iter=num_iter\n",
    "\n",
    "    def compute(self, images, labels):\n",
    "        criterion = nn.CrossEntropyLoss()\n",
    "        images = images.to(device)\n",
    "        labels = labels.to(device)\n",
    "        for t in range(self.num_iter):\n",
    "            images.requires_grad = True\n",
    "            outputs = self.model(images)\n",
    "            self.model.zero_grad()\n",
    "            loss = criterion(outputs, labels)\n",
    "            loss.backward()\n",
    "            adv_images = images + self.alpha*images.grad.sign()\n",
    "            eta = torch.clamp(adv_images - images.data, min=-self.eps, max=self.eps)\n",
    "            images = torch.clamp(images.data + eta, min=0, max=1).detach()\n",
    "        return images\n",
    "            "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "zDiS4YGxTF7F",
   "metadata": {
    "id": "zDiS4YGxTF7F",
    "outputId": "f59b0a62-b3b6-4f51-9b34-b8a2a91a0cf2"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Eval on attacked images :\n",
      "Correct : 0\n",
      "Total : 10000\n",
      "Accuracy: 0 %\n",
      "__________________________\n",
      "Eval on attacked and noised images :\n",
      "Correct : 485\n",
      "Total : 10000\n",
      "Accuracy: 4 %\n"
     ]
    }
   ],
   "source": [
    "model = VGG(\"VGG16\")\n",
    "model.load_state_dict(torch.load('./vgg16.pth'))\n",
    "model = model.to(device)\n",
    "\n",
    "attack = ProjectedGradientDescent(model,eps=0.03,alpha=0.007,num_iter=20)\n",
    "print(\"Eval on attacked images :\")\n",
    "eval_model(model, test_loader, attack)\n",
    "\n",
    "print(\"__________________________\")\n",
    "print(\"Eval on attacked and noised images :\")\n",
    "eval_model(model, test_loader, attack, noise=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b1YtzP0WTKs8",
   "metadata": {
    "id": "b1YtzP0WTKs8"
   },
   "source": [
    "Catastrophe ! Notre modèle est complement destabilisé par cette attaque. Il a même une pire accuracy que le classifieur random.\n",
    "\n",
    "Dans la suite de ce notebook nous verrons deux façon de rendre notre modèle robuste face aux attaques.\n",
    "\n",
    "- La data augmentation\n",
    "- L'adversarial training"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "53W7bvetb3jn",
   "metadata": {
    "id": "53W7bvetb3jn"
   },
   "source": [
    "#Data Augmentation"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "0jvGHsGjtl2b",
   "metadata": {
    "id": "0jvGHsGjtl2b"
   },
   "source": [
    "##Création de variante d'image"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9eojNh4OcEtn",
   "metadata": {
    "id": "9eojNh4OcEtn"
   },
   "source": [
    "L'augmentation des données consiste à créer de fausses données. Pour ce faire on utilise les images existantes pour en créer des variantes. \n",
    "Cela a deux effets :\n",
    "- Augmenter la taille de notre dataset.\n",
    "\n",
    "- Régulariser le réseau.\n",
    "\n",
    "\n",
    "Dans notre cas nous allons utiliser deux transformations:\n",
    "- RandomCrop: On ajoute une bordure à l'image et on découpe une nouvelle image de la taille souhaitée à partir de l'image avec bordure. Dans notre cas, pour une image 32*32, on ajoute une bordure de 4 pixels. L'image fait donc 40*40. Dans cette dernière on selectionne ensuite une image aléatoire de 32*32.\n",
    "\n",
    "![1_BX28QphEtCl0ZOofp6Bftg.png](data:image/png;base64,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)\n",
    "\n",
    "- RandomHorizontalFilp: Pour retourner l'image horizontalement.\n",
    "\n",
    "![1_NBYc-cg8TDOmc9AFLC_XbQ.png](data:image/png;base64,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)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "BkevJR6HcCmf",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "BkevJR6HcCmf",
    "outputId": "729a654e-7a55-42d9-a3e7-f6962a43d07b"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Files already downloaded and verified\n",
      "Files already downloaded and verified\n"
     ]
    }
   ],
   "source": [
    "transform_train = transforms.Compose([\n",
    "    transforms.RandomCrop(32, padding=4),\n",
    "    transforms.RandomHorizontalFlip(),\n",
    "    transforms.ToTensor(),\n",
    "    transforms.Normalize((0.4914, 0.4822, 0.4465), (0.2023, 0.1994, 0.2010)),\n",
    "])\n",
    "\n",
    "transform_test = transforms.Compose([\n",
    "    transforms.ToTensor(),\n",
    "    transforms.Normalize((0.4914, 0.4822, 0.4465), (0.2023, 0.1994, 0.2010)),\n",
    "])\n",
    "\n",
    "batch_size = 64\n",
    "\n",
    "trainset = torchvision.datasets.CIFAR10(\n",
    "    root='./data', train=True, download=True, transform=transform_train)\n",
    "train_loader = torch.utils.data.DataLoader(\n",
    "    trainset, batch_size=batch_size, shuffle=True, num_workers=2)\n",
    "\n",
    "testset = torchvision.datasets.CIFAR10(\n",
    "    root='./data', train=False, download=True, transform=transform_test)\n",
    "test_loader = torch.utils.data.DataLoader(\n",
    "    testset, batch_size=100, shuffle=False, num_workers=2)\n",
    "\n",
    "classes = ('plane', 'car', 'bird', 'cat', 'deer',\n",
    "           'dog', 'frog', 'horse', 'ship', 'truck')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "RCgBkloXts2e",
   "metadata": {
    "id": "RCgBkloXts2e"
   },
   "source": [
    "##Ajout de bruits "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "91dbeb38",
   "metadata": {
    "id": "91dbeb38"
   },
   "source": [
    "Pourquoi ne pas utiliser le bruit comme un outil pour de la data augmentation ?\n",
    "\n",
    "On obtient alors plus de données d’entraînement pour le modèle.\n",
    "On entraîne le modèle sur des données bruitées ce qui permet une meilleure généralisation sur les données bruitées. généralisera bien sur des données bruyantes également.\n",
    "Dans notre cas on va utiliser trois types de bruits:\n",
    "\n",
    "gaussian noise:\n",
    "\n",
    "salt and pepper noise:\n",
    "\n",
    "speckle noise:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "85fa4664",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 554
    },
    "id": "85fa4664",
    "outputId": "13b18c2e-9179-4311-9a41-32e677baa5b1"
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Clean images :\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Gaussian noise :\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Speckle noise :\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Salt&Pepper noise :\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "def add_gaussian_noise(images, seed=None):\n",
    "    images = images.to(device)\n",
    "    rng = np.random.default_rng(seed)\n",
    "    noise = torch.tensor(rng.normal(0, 0.05 ** 0.5, images.shape), dtype=torch.float32).to(device)\n",
    "    return   images + noise\n",
    "    \n",
    "\n",
    "def add_salt_pepper_noise(images):\n",
    "    images = torch.tensor(random_noise(images, mode='s&p', amount = 0.005, salt_vs_pepper = 0., clip=True), dtype=torch.float32).to(device)\n",
    "    return images\n",
    "\n",
    "def add_speckle_noise(images, seed=None):\n",
    "    images = images.to(device)\n",
    "    rng = np.random.default_rng(seed)\n",
    "    noise = torch.tensor(rng.normal(0, 0.05 ** 0.5, images.shape),  dtype=torch.float32).to(device)\n",
    "    return images + images * noise\n",
    "\n",
    "\n",
    "def add_noise(images):\n",
    "    noises = [\"gaussian\", \"speckle\", \"none\"]\n",
    "    noise = random.choice(noises)\n",
    "    if(noise == \"gaussian\"):\n",
    "        return add_gaussian_noise(images)\n",
    "    elif(noise == \"s&p\"):\n",
    "        return add_salt_pepper_noise(images)\n",
    "    elif(noise == \"s&p\"):\n",
    "        return add_speckle_noise(images)\n",
    "    return images\n",
    "\n",
    "\n",
    "num_img = 5\n",
    "example_loader = torch.utils.data.DataLoader(trainset, batch_size=num_img, shuffle=True, num_workers=2)\n",
    "dataiter = iter(example_loader)\n",
    "images, labels = dataiter.next()\n",
    "\n",
    "print(\"Clean images :\")\n",
    "imshow(torchvision.utils.make_grid(images))\n",
    "print(\"Gaussian noise :\")\n",
    "images_gaussian = add_gaussian_noise(images)\n",
    "imshow(torchvision.utils.make_grid(images_gaussian.cpu()))\n",
    "print(\"Speckle noise :\")\n",
    "images_speckle = add_speckle_noise(images)\n",
    "imshow(torchvision.utils.make_grid(images_speckle.cpu()))\n",
    "print(\"Salt&Pepper noise :\")\n",
    "images_sp = add_salt_pepper_noise(images)\n",
    "imshow(torchvision.utils.make_grid(images_sp.cpu()))\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "RF_Ee6pBUbKO",
   "metadata": {
    "id": "RF_Ee6pBUbKO"
   },
   "source": [
    "#Adversarial training"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "0DJxnpLFU5BT",
   "metadata": {
    "id": "0DJxnpLFU5BT"
   },
   "source": [
    "Nous avons vu avec les resultats précédents que notre modèle n'a pas su classifié les images attaqué avec le PGD. L'idée derrière l'adversarial training est d'attaquer les images d'entrainement et de s'entrainer sur ces dernières. Concrétement cela revient à rajouter une ligne dans notre fonction *training*:\n",
    "\n",
    "\n",
    "```\n",
    "images = attack.compute(images, labels)\n",
    "```\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "68fd5460",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 444
    },
    "id": "68fd5460",
    "outputId": "0cbd04d3-2c2c-4a46-8d99-294b47a63734"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training with attack :\n",
      "epoch 1 -> training_loss: 2.204\n",
      "epoch 2 -> training_loss: 2.120\n",
      "epoch 3 -> training_loss: 2.083\n",
      "epoch 4 -> training_loss: 2.051\n",
      "epoch 5 -> training_loss: 2.028\n",
      "epoch 6 -> training_loss: 1.997\n",
      "epoch 7 -> training_loss: 1.973\n",
      "epoch 8 -> training_loss: 1.945\n",
      "epoch 9 -> training_loss: 1.921\n",
      "epoch 10 -> training_loss: 1.899\n",
      "epoch 11 -> training_loss: 1.878\n",
      "epoch 12 -> training_loss: 1.856\n",
      "epoch 13 -> training_loss: 1.844\n",
      "epoch 14 -> training_loss: 1.825\n",
      "epoch 15 -> training_loss: 1.810\n",
      "epoch 16 -> training_loss: 1.793\n",
      "epoch 17 -> training_loss: 1.774\n",
      "epoch 18 -> training_loss: 1.760\n",
      "epoch 19 -> training_loss: 1.748\n",
      "epoch 20 -> training_loss: 1.728\n",
      "epoch 21 -> training_loss: 1.717\n",
      "epoch 22 -> training_loss: 1.704\n",
      "epoch 23 -> training_loss: 1.690\n",
      "epoch 24 -> training_loss: 1.675\n",
      "epoch 25 -> training_loss: 1.670\n",
      "epoch 26 -> training_loss: 1.654\n",
      "epoch 27 -> training_loss: 1.645\n",
      "epoch 28 -> training_loss: 1.636\n",
      "epoch 29 -> training_loss: 1.625\n",
      "epoch 30 -> training_loss: 1.621\n",
      "epoch 31 -> training_loss: 1.610\n",
      "epoch 32 -> training_loss: 1.598\n",
      "epoch 33 -> training_loss: 1.587\n",
      "epoch 34 -> training_loss: 1.582\n",
      "epoch 35 -> training_loss: 1.570\n",
      "epoch 36 -> training_loss: 1.565\n",
      "epoch 37 -> training_loss: 1.554\n",
      "epoch 38 -> training_loss: 1.547\n",
      "epoch 39 -> training_loss: 1.543\n",
      "epoch 40 -> training_loss: 1.531\n",
      "epoch 41 -> training_loss: 1.524\n",
      "epoch 42 -> training_loss: 1.521\n",
      "epoch 43 -> training_loss: 1.513\n",
      "epoch 44 -> training_loss: 1.510\n",
      "epoch 45 -> training_loss: 1.499\n",
      "epoch 46 -> training_loss: 1.494\n",
      "epoch 47 -> training_loss: 1.490\n",
      "epoch 48 -> training_loss: 1.482\n",
      "epoch 49 -> training_loss: 1.474\n",
      "epoch 50 -> training_loss: 1.470\n",
      "epoch 51 -> training_loss: 1.466\n",
      "epoch 52 -> training_loss: 1.463\n",
      "epoch 53 -> training_loss: 1.452\n",
      "epoch 54 -> training_loss: 1.444\n",
      "epoch 55 -> training_loss: 1.436\n",
      "epoch 56 -> training_loss: 1.431\n",
      "epoch 57 -> training_loss: 1.431\n",
      "epoch 58 -> training_loss: 1.428\n",
      "epoch 59 -> training_loss: 1.420\n",
      "epoch 60 -> training_loss: 1.413\n",
      "epoch 61 -> training_loss: 1.408\n",
      "epoch 62 -> training_loss: 1.403\n",
      "epoch 63 -> training_loss: 1.396\n",
      "epoch 64 -> training_loss: 1.387\n",
      "epoch 65 -> training_loss: 1.384\n",
      "epoch 66 -> training_loss: 1.381\n",
      "epoch 67 -> training_loss: 1.378\n",
      "epoch 68 -> training_loss: 1.368\n",
      "epoch 69 -> training_loss: 1.365\n",
      "epoch 70 -> training_loss: 1.359\n",
      "epoch 71 -> training_loss: 1.351\n",
      "epoch 72 -> training_loss: 1.346\n",
      "epoch 73 -> training_loss: 1.337\n",
      "epoch 74 -> training_loss: 1.332\n",
      "epoch 75 -> training_loss: 1.327\n",
      "epoch 76 -> training_loss: 1.321\n",
      "epoch 77 -> training_loss: 1.307\n",
      "epoch 78 -> training_loss: 1.308\n",
      "epoch 79 -> training_loss: 1.301\n",
      "epoch 80 -> training_loss: 1.298\n",
      "epoch 81 -> training_loss: 1.293\n",
      "epoch 82 -> training_loss: 1.284\n",
      "epoch 83 -> training_loss: 1.282\n",
      "epoch 84 -> training_loss: 1.276\n",
      "epoch 85 -> training_loss: 1.267\n",
      "epoch 86 -> training_loss: 1.256\n",
      "epoch 87 -> training_loss: 1.254\n",
      "epoch 88 -> training_loss: 1.247\n",
      "epoch 89 -> training_loss: 1.240\n",
      "epoch 90 -> training_loss: 1.232\n",
      "epoch 91 -> training_loss: 1.230\n",
      "epoch 92 -> training_loss: 1.224\n",
      "epoch 93 -> training_loss: 1.211\n",
      "epoch 94 -> training_loss: 1.208\n",
      "epoch 95 -> training_loss: 1.203\n",
      "epoch 96 -> training_loss: 1.200\n",
      "epoch 97 -> training_loss: 1.189\n",
      "epoch 98 -> training_loss: 1.178\n",
      "epoch 99 -> training_loss: 1.171\n",
      "epoch 100 -> training_loss: 1.174\n",
      "epoch 101 -> training_loss: 1.153\n",
      "epoch 102 -> training_loss: 1.159\n",
      "epoch 103 -> training_loss: 1.147\n",
      "epoch 104 -> training_loss: 1.138\n",
      "epoch 105 -> training_loss: 1.128\n",
      "epoch 106 -> training_loss: 1.127\n",
      "epoch 107 -> training_loss: 1.115\n",
      "epoch 108 -> training_loss: 1.108\n",
      "epoch 109 -> training_loss: 1.106\n",
      "epoch 110 -> training_loss: 1.100\n",
      "epoch 111 -> training_loss: 1.089\n",
      "epoch 112 -> training_loss: 1.077\n",
      "epoch 113 -> training_loss: 1.069\n",
      "epoch 114 -> training_loss: 1.064\n",
      "epoch 115 -> training_loss: 1.051\n",
      "epoch 116 -> training_loss: 1.046\n",
      "epoch 117 -> training_loss: 1.039\n",
      "epoch 118 -> training_loss: 1.028\n",
      "epoch 119 -> training_loss: 1.021\n",
      "epoch 120 -> training_loss: 1.016\n",
      "epoch 121 -> training_loss: 1.010\n",
      "epoch 122 -> training_loss: 0.992\n",
      "epoch 123 -> training_loss: 0.985\n",
      "epoch 124 -> training_loss: 0.980\n",
      "epoch 125 -> training_loss: 0.971\n",
      "epoch 126 -> training_loss: 0.961\n",
      "epoch 127 -> training_loss: 0.951\n",
      "epoch 128 -> training_loss: 0.952\n",
      "epoch 129 -> training_loss: 0.936\n",
      "epoch 130 -> training_loss: 0.923\n",
      "epoch 131 -> training_loss: 0.915\n",
      "epoch 132 -> training_loss: 0.903\n",
      "epoch 133 -> training_loss: 0.891\n",
      "epoch 134 -> training_loss: 0.884\n",
      "epoch 135 -> training_loss: 0.881\n",
      "epoch 136 -> training_loss: 0.867\n",
      "epoch 137 -> training_loss: 0.853\n",
      "epoch 138 -> training_loss: 0.839\n",
      "epoch 139 -> training_loss: 0.836\n",
      "epoch 140 -> training_loss: 0.822\n",
      "epoch 141 -> training_loss: 0.815\n",
      "epoch 142 -> training_loss: 0.802\n",
      "epoch 143 -> training_loss: 0.794\n",
      "epoch 144 -> training_loss: 0.780\n",
      "epoch 145 -> training_loss: 0.765\n",
      "epoch 146 -> training_loss: 0.767\n",
      "epoch 147 -> training_loss: 0.747\n",
      "epoch 148 -> training_loss: 0.742\n",
      "epoch 149 -> training_loss: 0.734\n",
      "epoch 150 -> training_loss: 0.721\n",
      "epoch 151 -> training_loss: 0.702\n",
      "epoch 152 -> training_loss: 0.699\n",
      "epoch 153 -> training_loss: 0.684\n",
      "epoch 154 -> training_loss: 0.672\n",
      "epoch 155 -> training_loss: 0.668\n",
      "epoch 156 -> training_loss: 0.655\n",
      "epoch 157 -> training_loss: 0.640\n",
      "epoch 158 -> training_loss: 0.634\n",
      "epoch 159 -> training_loss: 0.625\n",
      "epoch 160 -> training_loss: 0.606\n",
      "epoch 161 -> training_loss: 0.602\n",
      "epoch 162 -> training_loss: 0.587\n",
      "epoch 163 -> training_loss: 0.579\n",
      "epoch 164 -> training_loss: 0.569\n",
      "epoch 165 -> training_loss: 0.551\n",
      "epoch 166 -> training_loss: 0.544\n",
      "epoch 167 -> training_loss: 0.535\n",
      "epoch 168 -> training_loss: 0.530\n",
      "epoch 169 -> training_loss: 0.518\n",
      "epoch 170 -> training_loss: 0.508\n",
      "epoch 171 -> training_loss: 0.498\n",
      "epoch 172 -> training_loss: 0.487\n",
      "epoch 173 -> training_loss: 0.481\n",
      "epoch 174 -> training_loss: 0.472\n",
      "epoch 175 -> training_loss: 0.467\n",
      "epoch 176 -> training_loss: 0.458\n",
      "epoch 177 -> training_loss: 0.450\n",
      "epoch 178 -> training_loss: 0.444\n",
      "epoch 179 -> training_loss: 0.434\n",
      "epoch 180 -> training_loss: 0.424\n",
      "epoch 181 -> training_loss: 0.416\n",
      "epoch 182 -> training_loss: 0.412\n",
      "epoch 183 -> training_loss: 0.412\n",
      "epoch 184 -> training_loss: 0.407\n",
      "epoch 185 -> training_loss: 0.402\n",
      "epoch 186 -> training_loss: 0.396\n",
      "epoch 187 -> training_loss: 0.393\n",
      "epoch 188 -> training_loss: 0.391\n",
      "epoch 189 -> training_loss: 0.386\n",
      "epoch 190 -> training_loss: 0.382\n",
      "epoch 191 -> training_loss: 0.381\n",
      "epoch 192 -> training_loss: 0.379\n",
      "epoch 193 -> training_loss: 0.377\n",
      "epoch 194 -> training_loss: 0.377\n",
      "epoch 195 -> training_loss: 0.373\n",
      "epoch 196 -> training_loss: 0.373\n",
      "epoch 197 -> training_loss: 0.376\n",
      "epoch 198 -> training_loss: 0.373\n",
      "epoch 199 -> training_loss: 0.367\n",
      "epoch 200 -> training_loss: 0.369\n",
      "epoch 201 -> training_loss: 0.369\n",
      "epoch 202 -> training_loss: 0.371\n",
      "epoch 203 -> training_loss: 0.370\n",
      "epoch 204 -> training_loss: 0.367\n",
      "epoch 205 -> training_loss: 0.368\n",
      "epoch 206 -> training_loss: 0.371\n",
      "epoch 207 -> training_loss: 0.372\n",
      "epoch 208 -> training_loss: 0.371\n",
      "epoch 209 -> training_loss: 0.371\n",
      "epoch 210 -> training_loss: 0.368\n",
      "epoch 211 -> training_loss: 0.370\n",
      "epoch 212 -> training_loss: 0.372\n",
      "epoch 213 -> training_loss: 0.371\n",
      "epoch 214 -> training_loss: 0.366\n",
      "epoch 215 -> training_loss: 0.370\n",
      "epoch 216 -> training_loss: 0.375\n",
      "epoch 217 -> training_loss: 0.375\n",
      "epoch 218 -> training_loss: 0.373\n",
      "epoch 219 -> training_loss: 0.378\n",
      "epoch 220 -> training_loss: 0.376\n",
      "Finished Training\n",
      "Wall time: 12h 21min 53s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "model = VGG('VGG16')\n",
    "model = model.to(device)\n",
    "\n",
    "criterion = nn.CrossEntropyLoss()\n",
    "optimizer = optim.SGD(model.parameters(), lr=0.01, momentum=0.9,  weight_decay=5e-4) #lr = 0.001 ne fait pas mieux\n",
    "scheduler = torch.optim.lr_scheduler.CosineAnnealingLR(optimizer, T_max=200)\n",
    "\n",
    "n_epochs = 220\n",
    "\n",
    "attack = ProjectedGradientDescent(model,eps=0.03,alpha=0.007,num_iter=20)\n",
    "\n",
    "print(\"Training with attack :\")\n",
    "training(model, criterion, optimizer, scheduler, n_epochs, attack)\n",
    "save_model(model, \"./vgg16_adv.pth\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "lrArWq90aL7i",
   "metadata": {
    "id": "lrArWq90aL7i"
   },
   "source": [
    "#Evaluation du modèle après adversarial training"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "89a45de8",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "89a45de8",
    "outputId": "5afe94cd-97d2-404d-a00d-98c8397debf3"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Eval on clean images :\n",
      "Correct : 7238\n",
      "Total : 10000\n",
      "Accuracy: 72 %\n",
      "---- Num iter attack = 10 -----\n",
      "__________________________\n",
      "Eval on noised images :\n",
      "Correct : 5266\n",
      "Total : 10000\n",
      "Accuracy: 52 %\n",
      "__________________________\n",
      "Eval on attacked and noised images :\n",
      "Correct : 5350\n",
      "Total : 10000\n",
      "Accuracy: 53 %\n",
      "---- Num iter attack = 20 -----\n",
      "__________________________\n",
      "Eval on noised images :\n",
      "Correct : 4106\n",
      "Total : 10000\n",
      "Accuracy: 41 %\n",
      "__________________________\n",
      "Eval on attacked and noised images :\n",
      "Correct : 4181\n",
      "Total : 10000\n",
      "Accuracy: 41 %\n"
     ]
    }
   ],
   "source": [
    "#Eval Model Train With Adv Attack\n",
    "model = VGG(\"VGG16\")\n",
    "load_model(model, './vgg16_adv.pth')\n",
    "model = model.to(device)\n",
    "\n",
    "print(\"Eval on clean images :\")\n",
    "eval_model(model, test_loader)\n",
    "\n",
    "print(\"---- Num iter attack = 10 -----\")\n",
    "attack = ProjectedGradientDescent(model,eps=0.03,alpha=0.007,num_iter=10)\n",
    "print(\"__________________________\")\n",
    "print(\"Eval on attacked images :\")\n",
    "eval_model(model, test_loader, attack)\n",
    "print(\"__________________________\")\n",
    "print(\"Eval on attacked and noised images :\")\n",
    "eval_model(model, test_loader, attack, noise=True)\n",
    "\n",
    "print(\"---- Num iter attack = 20 -----\")\n",
    "attack = ProjectedGradientDescent(model,eps=0.03,alpha=0.007,num_iter=20)\n",
    "print(\"__________________________\")\n",
    "print(\"Eval on attacked images :\")\n",
    "eval_model(model, test_loader, attack)\n",
    "print(\"__________________________\")\n",
    "print(\"Eval on attacked and noised images :\")\n",
    "eval_model(model, test_loader, attack, noise=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "V2EL2xI5a7W5",
   "metadata": {
    "id": "V2EL2xI5a7W5"
   },
   "source": [
    "On voit que le modèle à perdu un peu en accuracy sur les images non attaquées mais en contrepartie il est beaucoup plus robuste sur les images attaquées."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "11lOUQeif4Gp",
   "metadata": {
    "id": "11lOUQeif4Gp"
   },
   "source": [
    "#Adversarial training + noise"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "443b9c40",
   "metadata": {
    "id": "443b9c40"
   },
   "source": [
    "On va maintenant entrainer un modèle en reprenant le principe d'adversarial training tout en rajoutant aléatoirement du bruit dans les images d'entrainement"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "e5d4a05e",
   "metadata": {
    "id": "e5d4a05e",
    "outputId": "c7a9d251-9054-4c2c-e9cf-08dbd681a6a7"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training with attack :\n",
      "epoch 1 -> training_loss: 2.196\n",
      "epoch 2 -> training_loss: 2.114\n",
      "epoch 3 -> training_loss: 2.076\n",
      "epoch 4 -> training_loss: 2.048\n",
      "epoch 5 -> training_loss: 2.026\n",
      "epoch 6 -> training_loss: 2.003\n",
      "epoch 7 -> training_loss: 1.971\n",
      "epoch 8 -> training_loss: 1.945\n",
      "epoch 9 -> training_loss: 1.925\n",
      "epoch 10 -> training_loss: 1.903\n",
      "epoch 11 -> training_loss: 1.885\n",
      "epoch 12 -> training_loss: 1.867\n",
      "epoch 13 -> training_loss: 1.852\n",
      "epoch 14 -> training_loss: 1.834\n",
      "epoch 15 -> training_loss: 1.819\n",
      "epoch 16 -> training_loss: 1.801\n",
      "epoch 17 -> training_loss: 1.785\n",
      "epoch 18 -> training_loss: 1.772\n",
      "epoch 19 -> training_loss: 1.762\n",
      "epoch 20 -> training_loss: 1.744\n",
      "epoch 21 -> training_loss: 1.737\n",
      "epoch 22 -> training_loss: 1.720\n",
      "epoch 23 -> training_loss: 1.707\n",
      "epoch 24 -> training_loss: 1.692\n",
      "epoch 25 -> training_loss: 1.683\n",
      "epoch 26 -> training_loss: 1.670\n",
      "epoch 27 -> training_loss: 1.666\n",
      "epoch 28 -> training_loss: 1.649\n",
      "epoch 29 -> training_loss: 1.641\n",
      "epoch 30 -> training_loss: 1.628\n",
      "epoch 31 -> training_loss: 1.625\n",
      "epoch 32 -> training_loss: 1.610\n",
      "epoch 33 -> training_loss: 1.604\n",
      "epoch 34 -> training_loss: 1.595\n",
      "epoch 35 -> training_loss: 1.588\n",
      "epoch 36 -> training_loss: 1.578\n",
      "epoch 37 -> training_loss: 1.575\n",
      "epoch 38 -> training_loss: 1.563\n",
      "epoch 39 -> training_loss: 1.556\n",
      "epoch 40 -> training_loss: 1.548\n",
      "epoch 41 -> training_loss: 1.539\n",
      "epoch 42 -> training_loss: 1.536\n",
      "epoch 43 -> training_loss: 1.529\n",
      "epoch 44 -> training_loss: 1.522\n",
      "epoch 45 -> training_loss: 1.513\n",
      "epoch 46 -> training_loss: 1.510\n",
      "epoch 47 -> training_loss: 1.501\n",
      "epoch 48 -> training_loss: 1.500\n",
      "epoch 49 -> training_loss: 1.493\n",
      "epoch 50 -> training_loss: 1.485\n",
      "epoch 51 -> training_loss: 1.480\n",
      "epoch 52 -> training_loss: 1.476\n",
      "epoch 53 -> training_loss: 1.469\n",
      "epoch 54 -> training_loss: 1.463\n",
      "epoch 55 -> training_loss: 1.457\n",
      "epoch 56 -> training_loss: 1.455\n",
      "epoch 57 -> training_loss: 1.446\n",
      "epoch 58 -> training_loss: 1.441\n",
      "epoch 59 -> training_loss: 1.438\n",
      "epoch 60 -> training_loss: 1.430\n",
      "epoch 61 -> training_loss: 1.426\n",
      "epoch 62 -> training_loss: 1.419\n",
      "epoch 63 -> training_loss: 1.414\n",
      "epoch 64 -> training_loss: 1.407\n",
      "epoch 65 -> training_loss: 1.406\n",
      "epoch 66 -> training_loss: 1.399\n",
      "epoch 67 -> training_loss: 1.391\n",
      "epoch 68 -> training_loss: 1.392\n",
      "epoch 69 -> training_loss: 1.386\n",
      "epoch 70 -> training_loss: 1.376\n",
      "epoch 71 -> training_loss: 1.367\n",
      "epoch 72 -> training_loss: 1.362\n",
      "epoch 73 -> training_loss: 1.354\n",
      "epoch 74 -> training_loss: 1.358\n",
      "epoch 75 -> training_loss: 1.345\n",
      "epoch 76 -> training_loss: 1.340\n",
      "epoch 77 -> training_loss: 1.335\n",
      "epoch 78 -> training_loss: 1.327\n",
      "epoch 79 -> training_loss: 1.325\n",
      "epoch 80 -> training_loss: 1.321\n",
      "epoch 81 -> training_loss: 1.314\n",
      "epoch 82 -> training_loss: 1.311\n",
      "epoch 83 -> training_loss: 1.302\n",
      "epoch 84 -> training_loss: 1.299\n",
      "epoch 85 -> training_loss: 1.290\n",
      "epoch 86 -> training_loss: 1.287\n",
      "epoch 87 -> training_loss: 1.277\n",
      "epoch 88 -> training_loss: 1.273\n",
      "epoch 89 -> training_loss: 1.260\n",
      "epoch 90 -> training_loss: 1.259\n",
      "epoch 91 -> training_loss: 1.254\n",
      "epoch 92 -> training_loss: 1.243\n",
      "epoch 93 -> training_loss: 1.244\n",
      "epoch 94 -> training_loss: 1.236\n",
      "epoch 95 -> training_loss: 1.229\n",
      "epoch 96 -> training_loss: 1.214\n",
      "epoch 97 -> training_loss: 1.212\n",
      "epoch 98 -> training_loss: 1.205\n",
      "epoch 99 -> training_loss: 1.199\n",
      "epoch 100 -> training_loss: 1.195\n",
      "epoch 101 -> training_loss: 1.187\n",
      "epoch 102 -> training_loss: 1.183\n",
      "epoch 103 -> training_loss: 1.169\n",
      "epoch 104 -> training_loss: 1.167\n",
      "epoch 105 -> training_loss: 1.167\n",
      "epoch 106 -> training_loss: 1.156\n",
      "epoch 107 -> training_loss: 1.141\n",
      "epoch 108 -> training_loss: 1.137\n",
      "epoch 109 -> training_loss: 1.130\n",
      "epoch 110 -> training_loss: 1.122\n",
      "epoch 111 -> training_loss: 1.114\n",
      "epoch 112 -> training_loss: 1.112\n",
      "epoch 113 -> training_loss: 1.099\n",
      "epoch 114 -> training_loss: 1.092\n",
      "epoch 115 -> training_loss: 1.079\n",
      "epoch 116 -> training_loss: 1.074\n",
      "epoch 117 -> training_loss: 1.067\n",
      "epoch 118 -> training_loss: 1.064\n",
      "epoch 119 -> training_loss: 1.050\n",
      "epoch 120 -> training_loss: 1.048\n",
      "epoch 121 -> training_loss: 1.040\n",
      "epoch 122 -> training_loss: 1.024\n",
      "epoch 123 -> training_loss: 1.014\n",
      "epoch 124 -> training_loss: 1.012\n",
      "epoch 125 -> training_loss: 1.002\n",
      "epoch 126 -> training_loss: 0.985\n",
      "epoch 127 -> training_loss: 0.986\n",
      "epoch 128 -> training_loss: 0.978\n",
      "epoch 129 -> training_loss: 0.962\n",
      "epoch 130 -> training_loss: 0.958\n",
      "epoch 131 -> training_loss: 0.945\n",
      "epoch 132 -> training_loss: 0.938\n",
      "epoch 133 -> training_loss: 0.925\n",
      "epoch 134 -> training_loss: 0.916\n",
      "epoch 135 -> training_loss: 0.912\n",
      "epoch 136 -> training_loss: 0.894\n",
      "epoch 137 -> training_loss: 0.887\n",
      "epoch 138 -> training_loss: 0.877\n",
      "epoch 139 -> training_loss: 0.863\n",
      "epoch 140 -> training_loss: 0.853\n",
      "epoch 141 -> training_loss: 0.843\n",
      "epoch 142 -> training_loss: 0.836\n",
      "epoch 143 -> training_loss: 0.830\n",
      "epoch 144 -> training_loss: 0.820\n",
      "epoch 145 -> training_loss: 0.804\n",
      "epoch 146 -> training_loss: 0.792\n",
      "epoch 147 -> training_loss: 0.783\n",
      "epoch 148 -> training_loss: 0.774\n",
      "epoch 149 -> training_loss: 0.766\n",
      "epoch 150 -> training_loss: 0.747\n",
      "epoch 151 -> training_loss: 0.737\n",
      "epoch 152 -> training_loss: 0.728\n",
      "epoch 153 -> training_loss: 0.719\n",
      "epoch 154 -> training_loss: 0.712\n",
      "epoch 155 -> training_loss: 0.697\n",
      "epoch 156 -> training_loss: 0.683\n",
      "epoch 157 -> training_loss: 0.675\n",
      "epoch 158 -> training_loss: 0.668\n",
      "epoch 159 -> training_loss: 0.652\n",
      "epoch 160 -> training_loss: 0.640\n",
      "epoch 161 -> training_loss: 0.622\n",
      "epoch 162 -> training_loss: 0.616\n",
      "epoch 163 -> training_loss: 0.608\n",
      "epoch 164 -> training_loss: 0.595\n",
      "epoch 165 -> training_loss: 0.585\n",
      "epoch 166 -> training_loss: 0.572\n",
      "epoch 167 -> training_loss: 0.563\n",
      "epoch 168 -> training_loss: 0.553\n",
      "epoch 169 -> training_loss: 0.544\n",
      "epoch 170 -> training_loss: 0.534\n",
      "epoch 171 -> training_loss: 0.528\n",
      "epoch 172 -> training_loss: 0.516\n",
      "epoch 173 -> training_loss: 0.504\n",
      "epoch 174 -> training_loss: 0.501\n",
      "epoch 175 -> training_loss: 0.489\n",
      "epoch 176 -> training_loss: 0.480\n",
      "epoch 177 -> training_loss: 0.472\n",
      "epoch 178 -> training_loss: 0.460\n",
      "epoch 179 -> training_loss: 0.461\n",
      "epoch 180 -> training_loss: 0.450\n",
      "epoch 181 -> training_loss: 0.444\n",
      "epoch 182 -> training_loss: 0.437\n",
      "epoch 183 -> training_loss: 0.435\n",
      "epoch 184 -> training_loss: 0.427\n",
      "epoch 185 -> training_loss: 0.422\n",
      "epoch 186 -> training_loss: 0.421\n",
      "epoch 187 -> training_loss: 0.416\n",
      "epoch 188 -> training_loss: 0.411\n",
      "epoch 189 -> training_loss: 0.409\n",
      "epoch 190 -> training_loss: 0.404\n",
      "epoch 191 -> training_loss: 0.402\n",
      "epoch 192 -> training_loss: 0.401\n",
      "epoch 193 -> training_loss: 0.397\n",
      "epoch 194 -> training_loss: 0.397\n",
      "epoch 195 -> training_loss: 0.395\n",
      "epoch 196 -> training_loss: 0.395\n",
      "epoch 197 -> training_loss: 0.391\n",
      "epoch 198 -> training_loss: 0.392\n",
      "epoch 199 -> training_loss: 0.392\n",
      "epoch 200 -> training_loss: 0.392\n",
      "epoch 201 -> training_loss: 0.391\n",
      "epoch 202 -> training_loss: 0.393\n",
      "epoch 203 -> training_loss: 0.393\n",
      "epoch 204 -> training_loss: 0.392\n",
      "epoch 205 -> training_loss: 0.391\n",
      "epoch 206 -> training_loss: 0.390\n",
      "epoch 207 -> training_loss: 0.390\n",
      "epoch 208 -> training_loss: 0.396\n",
      "epoch 209 -> training_loss: 0.391\n",
      "epoch 210 -> training_loss: 0.390\n",
      "epoch 211 -> training_loss: 0.391\n",
      "epoch 212 -> training_loss: 0.392\n",
      "epoch 213 -> training_loss: 0.389\n",
      "epoch 214 -> training_loss: 0.391\n",
      "epoch 215 -> training_loss: 0.393\n",
      "epoch 216 -> training_loss: 0.394\n",
      "epoch 217 -> training_loss: 0.394\n",
      "epoch 218 -> training_loss: 0.397\n",
      "epoch 219 -> training_loss: 0.396\n",
      "epoch 220 -> training_loss: 0.399\n",
      "Finished Training\n",
      "Wall time: 12h 18min 17s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "model = VGG('VGG16')\n",
    "model = model.to(device)\n",
    "\n",
    "criterion = nn.CrossEntropyLoss()\n",
    "optimizer = optim.SGD(model.parameters(), lr=0.01, momentum=0.9,  weight_decay=5e-4)\n",
    "scheduler = torch.optim.lr_scheduler.CosineAnnealingLR(optimizer, T_max=200)\n",
    "\n",
    "n_epochs = 220\n",
    "\n",
    "attack = ProjectedGradientDescent(model,eps=0.03,alpha=0.007,num_iter=20)\n",
    "\n",
    "print(\"Training with attack :\")\n",
    "training(model, criterion, optimizer, scheduler, n_epochs, attack, noise=True)\n",
    "save_model(model, \"./vgg16_adv_noise.pth\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "d6bdcbca",
   "metadata": {
    "id": "d6bdcbca",
    "outputId": "12418642-0ea6-4953-95d0-0a3588d3b53f"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Eval on clean images :\n",
      "Correct : 7258\n",
      "Total : 10000\n",
      "Accuracy: 72 %\n",
      "---- Num iter attack = 10 -----\n",
      "__________________________\n",
      "Eval on attacked images :\n",
      "Correct : 5195\n",
      "Total : 10000\n",
      "Accuracy: 51 %\n",
      "__________________________\n",
      "Eval on attacked and noised images :\n",
      "Correct : 5318\n",
      "Total : 10000\n",
      "Accuracy: 53 %\n",
      "---- Num iter attack = 20 -----\n",
      "__________________________\n",
      "Eval on attacked images :\n",
      "Correct : 4030\n",
      "Total : 10000\n",
      "Accuracy: 40 %\n",
      "__________________________\n",
      "Eval on attacked and noised images :\n",
      "Correct : 4112\n",
      "Total : 10000\n",
      "Accuracy: 41 %\n"
     ]
    }
   ],
   "source": [
    "#Eval Model Train With Adv Attack and noise\n",
    "model = VGG(\"VGG16\")\n",
    "load_model(model, './vgg16_adv_noise.pth')\n",
    "model = model.to(device)\n",
    "\n",
    "print(\"Eval on clean images :\")\n",
    "eval_model(model, test_loader)\n",
    "\n",
    "print(\"---- Num iter attack = 10 -----\")\n",
    "attack = ProjectedGradientDescent(model,eps=0.03,alpha=0.007,num_iter=10)\n",
    "print(\"__________________________\")\n",
    "print(\"Eval on attacked images :\")\n",
    "eval_model(model, test_loader, attack)\n",
    "print(\"__________________________\")\n",
    "print(\"Eval on attacked and noised images :\")\n",
    "eval_model(model, test_loader, attack, noise=True)\n",
    "\n",
    "print(\"---- Num iter attack = 20 -----\")\n",
    "attack = ProjectedGradientDescent(model,eps=0.03,alpha=0.007,num_iter=20)\n",
    "print(\"__________________________\")\n",
    "print(\"Eval on attacked images :\")\n",
    "eval_model(model, test_loader, attack)\n",
    "print(\"__________________________\")\n",
    "print(\"Eval on attacked and noised images :\")\n",
    "eval_model(model, test_loader, attack, noise=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "dda739af",
   "metadata": {
    "id": "dda739af"
   },
   "source": [
    "On voit que les accuracy sont similaires au premier modèle entrainé avec l'adversarial training. Cela est peut être du au fait que nous n'avons pas assez mis de noise dans nos images."
   ]
  }
 ],
 "metadata": {
  "accelerator": "GPU",
  "colab": {
   "collapsed_sections": [
    "yGh92aFCArud"
   ],
   "name": "PSD_final.ipynb",
   "provenance": []
  },
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "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": "3.9.7"
  },
  "widgets": {
   "application/vnd.jupyter.widget-state+json": {
    "0791008ba84b432492c984cb0cec8422": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "HBoxModel",
     "state": {
      "_dom_classes": [],
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "HBoxModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/controls",
      "_view_module_version": "1.5.0",
      "_view_name": "HBoxView",
      "box_style": "",
      "children": [
       "IPY_MODEL_70752c805d8f45e0a52db2636a321319",
       "IPY_MODEL_79a343f676314d5ba3a5b2df9422595a",
       "IPY_MODEL_0dda7b18e0204c71927d29e3996553ae"
      ],
      "layout": "IPY_MODEL_5d6a358b29d646c2a1e2bf80b4a7b9fc"
     }
    },
    "0892224ca7b94e42a77b4427d40f354e": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "DescriptionStyleModel",
     "state": {
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "DescriptionStyleModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "StyleView",
      "description_width": ""
     }
    },
    "0dda7b18e0204c71927d29e3996553ae": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "HTMLModel",
     "state": {
      "_dom_classes": [],
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "HTMLModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/controls",
      "_view_module_version": "1.5.0",
      "_view_name": "HTMLView",
      "description": "",
      "description_tooltip": null,
      "layout": "IPY_MODEL_bb1791ae91fc48eca9d628a26daf7f84",
      "placeholder": "​",
      "style": "IPY_MODEL_0892224ca7b94e42a77b4427d40f354e",
      "value": " 170499072/? [00:03&lt;00:00, 49800325.64it/s]"
     }
    },
    "182d6ff5143543218e36f5e4e28de773": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "ProgressStyleModel",
     "state": {
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "ProgressStyleModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "StyleView",
      "bar_color": null,
      "description_width": ""
     }
    },
    "5d6a358b29d646c2a1e2bf80b4a7b9fc": {
     "model_module": "@jupyter-widgets/base",
     "model_module_version": "1.2.0",
     "model_name": "LayoutModel",
     "state": {
      "_model_module": "@jupyter-widgets/base",
      "_model_module_version": "1.2.0",
      "_model_name": "LayoutModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "LayoutView",
      "align_content": null,
      "align_items": null,
      "align_self": null,
      "border": null,
      "bottom": null,
      "display": null,
      "flex": null,
      "flex_flow": null,
      "grid_area": null,
      "grid_auto_columns": null,
      "grid_auto_flow": null,
      "grid_auto_rows": null,
      "grid_column": null,
      "grid_gap": null,
      "grid_row": null,
      "grid_template_areas": null,
      "grid_template_columns": null,
      "grid_template_rows": null,
      "height": null,
      "justify_content": null,
      "justify_items": null,
      "left": null,
      "margin": null,
      "max_height": null,
      "max_width": null,
      "min_height": null,
      "min_width": null,
      "object_fit": null,
      "object_position": null,
      "order": null,
      "overflow": null,
      "overflow_x": null,
      "overflow_y": null,
      "padding": null,
      "right": null,
      "top": null,
      "visibility": null,
      "width": null
     }
    },
    "70752c805d8f45e0a52db2636a321319": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "HTMLModel",
     "state": {
      "_dom_classes": [],
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "HTMLModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/controls",
      "_view_module_version": "1.5.0",
      "_view_name": "HTMLView",
      "description": "",
      "description_tooltip": null,
      "layout": "IPY_MODEL_daa90939e369449c92587365fd0c3490",
      "placeholder": "​",
      "style": "IPY_MODEL_9a03fd5d95614d86b1e501282c04c679",
      "value": ""
     }
    },
    "79a343f676314d5ba3a5b2df9422595a": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "FloatProgressModel",
     "state": {
      "_dom_classes": [],
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "FloatProgressModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/controls",
      "_view_module_version": "1.5.0",
      "_view_name": "ProgressView",
      "bar_style": "success",
      "description": "",
      "description_tooltip": null,
      "layout": "IPY_MODEL_f0e8828a28294f3d9c444b7c948e8f3d",
      "max": 170498071,
      "min": 0,
      "orientation": "horizontal",
      "style": "IPY_MODEL_182d6ff5143543218e36f5e4e28de773",
      "value": 170498071
     }
    },
    "9a03fd5d95614d86b1e501282c04c679": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "DescriptionStyleModel",
     "state": {
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "DescriptionStyleModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "StyleView",
      "description_width": ""
     }
    },
    "bb1791ae91fc48eca9d628a26daf7f84": {
     "model_module": "@jupyter-widgets/base",
     "model_module_version": "1.2.0",
     "model_name": "LayoutModel",
     "state": {
      "_model_module": "@jupyter-widgets/base",
      "_model_module_version": "1.2.0",
      "_model_name": "LayoutModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "LayoutView",
      "align_content": null,
      "align_items": null,
      "align_self": null,
      "border": null,
      "bottom": null,
      "display": null,
      "flex": null,
      "flex_flow": null,
      "grid_area": null,
      "grid_auto_columns": null,
      "grid_auto_flow": null,
      "grid_auto_rows": null,
      "grid_column": null,
      "grid_gap": null,
      "grid_row": null,
      "grid_template_areas": null,
      "grid_template_columns": null,
      "grid_template_rows": null,
      "height": null,
      "justify_content": null,
      "justify_items": null,
      "left": null,
      "margin": null,
      "max_height": null,
      "max_width": null,
      "min_height": null,
      "min_width": null,
      "object_fit": null,
      "object_position": null,
      "order": null,
      "overflow": null,
      "overflow_x": null,
      "overflow_y": null,
      "padding": null,
      "right": null,
      "top": null,
      "visibility": null,
      "width": null
     }
    },
    "daa90939e369449c92587365fd0c3490": {
     "model_module": "@jupyter-widgets/base",
     "model_module_version": "1.2.0",
     "model_name": "LayoutModel",
     "state": {
      "_model_module": "@jupyter-widgets/base",
      "_model_module_version": "1.2.0",
      "_model_name": "LayoutModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "LayoutView",
      "align_content": null,
      "align_items": null,
      "align_self": null,
      "border": null,
      "bottom": null,
      "display": null,
      "flex": null,
      "flex_flow": null,
      "grid_area": null,
      "grid_auto_columns": null,
      "grid_auto_flow": null,
      "grid_auto_rows": null,
      "grid_column": null,
      "grid_gap": null,
      "grid_row": null,
      "grid_template_areas": null,
      "grid_template_columns": null,
      "grid_template_rows": null,
      "height": null,
      "justify_content": null,
      "justify_items": null,
      "left": null,
      "margin": null,
      "max_height": null,
      "max_width": null,
      "min_height": null,
      "min_width": null,
      "object_fit": null,
      "object_position": null,
      "order": null,
      "overflow": null,
      "overflow_x": null,
      "overflow_y": null,
      "padding": null,
      "right": null,
      "top": null,
      "visibility": null,
      "width": null
     }
    },
    "f0e8828a28294f3d9c444b7c948e8f3d": {
     "model_module": "@jupyter-widgets/base",
     "model_module_version": "1.2.0",
     "model_name": "LayoutModel",
     "state": {
      "_model_module": "@jupyter-widgets/base",
      "_model_module_version": "1.2.0",
      "_model_name": "LayoutModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "LayoutView",
      "align_content": null,
      "align_items": null,
      "align_self": null,
      "border": null,
      "bottom": null,
      "display": null,
      "flex": null,
      "flex_flow": null,
      "grid_area": null,
      "grid_auto_columns": null,
      "grid_auto_flow": null,
      "grid_auto_rows": null,
      "grid_column": null,
      "grid_gap": null,
      "grid_row": null,
      "grid_template_areas": null,
      "grid_template_columns": null,
      "grid_template_rows": null,
      "height": null,
      "justify_content": null,
      "justify_items": null,
      "left": null,
      "margin": null,
      "max_height": null,
      "max_width": null,
      "min_height": null,
      "min_width": null,
      "object_fit": null,
      "object_position": null,
      "order": null,
      "overflow": null,
      "overflow_x": null,
      "overflow_y": null,
      "padding": null,
      "right": null,
      "top": null,
      "visibility": null,
      "width": null
     }
    }
   }
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}