White_Box_Attack_PyTorch.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "White-Box-Attack_PyTorch",
      "provenance": []
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "code",
      "metadata": {
        "id": "628Hla1CJ8do",
        "colab_type": "code",
        "outputId": "fb36dfc6-49dd-4e45-d87a-30a2ee8232b0",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 126
        }
      },
      "source": [
        "import sys\n",
        "import pandas as pd\n",
        "import numpy as np\n",
        "import matplotlib.pyplot as plt\n",
        "import torch\n",
        "import torch.nn as nn\n",
        "import torch.nn.functional as F\n",
        "from torch import optim\n",
        "from torch.utils.data import TensorDataset, DataLoader, Dataset\n",
        "from torchvision import datasets, transforms\n",
        "from skimage.transform import rotate\n",
        "from sklearn.model_selection import train_test_split\n",
        "from sklearn.metrics import accuracy_score\n",
        "from google.colab import drive \n",
        "drive.mount('/content/drive')"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Go to this URL in a browser: https://accounts.google.com/o/oauth2/auth?client_id=947318989803-6bn6qk8qdgf4n4g3pfee6491hc0brc4i.apps.googleusercontent.com&redirect_uri=urn%3aietf%3awg%3aoauth%3a2.0%3aoob&response_type=code&scope=email%20https%3a%2f%2fwww.googleapis.com%2fauth%2fdocs.test%20https%3a%2f%2fwww.googleapis.com%2fauth%2fdrive%20https%3a%2f%2fwww.googleapis.com%2fauth%2fdrive.photos.readonly%20https%3a%2f%2fwww.googleapis.com%2fauth%2fpeopleapi.readonly\n",
            "\n",
            "Enter your authorization code:\n",
            "··········\n",
            "Mounted at /content/drive\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "mwSiSTFJKNYD",
        "colab_type": "code",
        "outputId": "7f787c5d-de23-4faa-c68c-2bd76779dede",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 70
        }
      },
      "source": [
        "print(sys.version)\n",
        "device='cuda'\n",
        "\n",
        "train_on_gpu = torch.cuda.is_available()\n",
        "\n",
        "if not train_on_gpu:\n",
        "    print('CUDA is not available.  Training on CPU ...')\n",
        "else:\n",
        "    print('CUDA is available!  Training on GPU ...')"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "3.6.9 (default, Apr 18 2020, 01:56:04) \n",
            "[GCC 8.4.0]\n",
            "CUDA is available!  Training on GPU ...\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "6XTfeJa5KQ_Z",
        "colab_type": "code",
        "outputId": "5fab7ef3-34a3-4188-93f9-49778eea8e1b",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 52
        }
      },
      "source": [
        "df=pd.read_csv(\"/content/drive/My Drive/Colab Notebooks/MNIST/train.csv\")\n",
        "train=df.values\n",
        "x_train=(train[:,1:]).astype(np.float32)\n",
        "y_train=(train[:,0])\n",
        "\n",
        "del df,train\n",
        "\n",
        "print(x_train.shape)\n",
        "print(y_train.shape)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "(42000, 784)\n",
            "(42000,)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "dPdoNuLIKkH_",
        "colab_type": "code",
        "outputId": "b03bd5bf-4de6-4571-b66c-c92899192f6d",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 106
        }
      },
      "source": [
        "random_seed = 1\n",
        "x_train, x_test, y_train, y_test = train_test_split(x_train, y_train, test_size = 7000, random_state = random_seed)\n",
        "\n",
        "print(x_train.shape)\n",
        "print(x_train.dtype)\n",
        "print(x_test.shape)\n",
        "print(y_train.shape)\n",
        "print(y_test.shape)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "(35000, 784)\n",
            "float32\n",
            "(7000, 784)\n",
            "(35000,)\n",
            "(7000,)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Kg6rw4-zRCtA",
        "colab_type": "code",
        "outputId": "141b6e35-6744-4f24-daa7-950742a5d279",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 52
        }
      },
      "source": [
        "X_train = []\n",
        "Y_train = []\n",
        "np.random.seed(1)\n",
        "rot_angle = 40*np.random.random_sample([x_train.shape[0],4])\n",
        "for i in range(x_train.shape[0]):\n",
        "    X_train.append(x_train[i])\n",
        "    for j in range(4):\n",
        "      X_train.append(rotate(x_train[i].reshape((28,28)),angle=rot_angle[i,j],mode='wrap').reshape((784))) \n",
        "    for j in range(5):\n",
        "      Y_train.append(y_train[i])\n",
        "\n",
        "X_train = np.asarray(X_train,dtype=np.float32)\n",
        "Y_train = np.asarray(Y_train)\n",
        "print(X_train.shape)\n",
        "print(Y_train.shape)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "(175000, 784)\n",
            "(175000,)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Df3lp-vSRECT",
        "colab_type": "code",
        "outputId": "fc45f7d0-e575-439a-d0ca-ce5ff37a48f4",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 301
        }
      },
      "source": [
        "index = 5\n",
        "print(Y_train[index])\n",
        "plt.imshow(X_train[index].reshape((28,28)), cmap='gray')"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "7\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.image.AxesImage at 0x7f8a43d1a4a8>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 6
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAPsAAAD4CAYAAAAq5pAIAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjEsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+j8jraAAANCElEQVR4nO3dX6xV9ZnG8edB2huoETRzgpQMWPWiaZCOBE0ktY2BKDdIYghcTCAxHi7qpI29KGEu6pWaSW1TbxqpmMKkY21sDcQ0M2UICcMNelAE1LRaAiknB84YYhA16QDvXJyFOeLZv31Ye+0/h/f7SU723uvda683Wx/W2uvfzxEhANe/Wf1uAEBvEHYgCcIOJEHYgSQIO5DE7F4uzDa7/oEuiwhPNb2jNbvtB23/2fYHtrd28lkAust1j7PbvkHSXyStknRa0huSNkbEu4V5WLMDXdaNNfsKSR9ExImI+Luk30pa28HnAeiiTsK+UNLfJr0+XU37AtvDtkdsj3SwLAAd6voOuojYLmm7xGY80E+drNlHJS2a9Prr1TQAA6iTsL8h6Q7bS2x/VdIGSXuaaQtA02pvxkfERduPS/ovSTdIejEi3mmsMwCNqn3ordbC+M0OdF1XTqoBMHMQdiAJwg4kQdiBJAg7kARhB5Ig7EAShB1IgrADSRB2IAnCDiRB2IEkCDuQBGEHkiDsQBKEHUiCsANJEHYgCcIOJEHYgSQIO5AEYQeSIOxAEoQdSIKwA0kQdiAJwg4kQdiBJAg7kARhB5KoPT67JNk+KeljSZckXYyI5U00BaB5HYW98r2I+LCBzwHQRWzGA0l0GvaQ9Cfbh20PT/UG28O2R2yPdLgsAB1wRNSf2V4YEaO2/0HSXkn/EhEHCu+vvzAA0xIRnmp6R2v2iBitHsclvSppRSefB6B7aofd9hzbX7vyXNJqScebagxAszrZGz8k6VXbVz7nPyLiPxvpCjPGrbfeWqwfONDyV51uu+224rz33ntvsf76668X6/ii2mGPiBOS7mqwFwBdxKE3IAnCDiRB2IEkCDuQBGEHkmjiQhhcx2666aZi/bHHHivWlyxZ0rLW7uzNbdu2FeuPPPJIsX7x4sViPRvW7EAShB1IgrADSRB2IAnCDiRB2IEkCDuQREd3qrnmhXV4p5pZs1r/29Tucsl2x1zHxsaK9blz57asffrpp8V5P/vss2J98eLFxfrs2fVPh2h3rLrdf/+VK1cW67fffvs19zRdFy5cKNaXLl1arJ86darJdmaMrtypBsDMQdiBJAg7kARhB5Ig7EAShB1IgrADScyo69lfeeWVlrW1a9cW5213zPbQoUPF+sKFC1vWzpw5U5z3ww/L414+9NBDxfqcOXOK9ZLqVt8t9fI8i2u1ZcuWYj3rcfS6WLMDSRB2IAnCDiRB2IEkCDuQBGEHkiDsQBIz6nr2y5cvt6wN8vHifjp48GCxvnfv3mJ9dHS0WH/hhReuuacrRkZGivVVq1YV6+fPn6+97OtZ7evZbb9oe9z28UnT5tvea/v96nFek80CaN50NuN/LenBq6ZtlbQvIu6QtK96DWCAtQ17RByQdO6qyWsl7aye75T0cMN9AWhY3XPjhyLiyk3bzkgaavVG28OShmsuB0BDOr4QJiKitOMtIrZL2i51voMOQH11D72dtb1AkqrH8eZaAtANdcO+R9Km6vkmSbubaQdAt7TdjLf9kqTvSrrF9mlJP5H0jKTf2X5U0ilJ67vZ5BUPPPBAy9oTTzxRnPejjz4q1p9//vlaPQ26dsfZ2+nmfeHHx8sbhBxHb1bbsEfExhal1skDMHA4XRZIgrADSRB2IAnCDiRB2IEkZtStpPfv31+rBoA1O5AGYQeSIOxAEoQdSIKwA0kQdiAJwg4kMaOOs1+v7rzzzmL95ZdfLtZLQzo//fTTtXq6Ymio5R3HOrZo0aJi/eabb+6oft9997WsHT16tDjv4cOHi/WZiDU7kARhB5Ig7EAShB1IgrADSRB2IAnCDiQxo4Zsvl699dZbxfrSpUtrf7Y95ei9nxvkoa7Pnbt6iMEvmj27fJrIjTfe2LL2ySef1J530NUeshnA9YGwA0kQdiAJwg4kQdiBJAg7kARhB5LgevYB8PbbbxfrnRxnn8nmz5/f0fzPPfdcy9qFCxc6+uyZqO2a3faLtsdtH5807Unbo7aPVH9rutsmgE5NZzP+15IenGL6zyNiWfX3x2bbAtC0tmGPiAOSyuctAhh4neyge9z20Wozf16rN9ketj1ie6SDZQHoUN2w/1LSNyQtkzQm6dlWb4yI7RGxPCKW11wWgAbUCntEnI2ISxFxWdKvJK1oti0ATasVdtsLJr1cJ+l4q/cCGAxtr2e3/ZKk70q6RdJZST+pXi+TFJJOStoSEWNtF8b17FO66667ivU1a+of2Wx33/jVq1cX60899VSxfvfddxfrpf+/du/eXZx33bp1xTqm1up69rYn1UTExikm7+i4IwA9xemyQBKEHUiCsANJEHYgCcIOJMGtpFG0Y0f5wMvmzZtrf/b9999frB88eLD2Z2fGraSB5Ag7kARhB5Ig7EAShB1IgrADSRB2IAnCDiRB2IEkCDuQBGEHkiDsQBKEHUiCsANJEHYgCcIOJEHYgSQIO5AEYQeSIOxAEoQdSIKwA0kQdiCJtqO4At2yfv36Yp37xjer7Zrd9iLb+22/a/sd2z+ops+3vdf2+9XjvO63C6Cu6WzGX5T0o4j4pqR7JX3f9jclbZW0LyLukLSveg1gQLUNe0SMRcSb1fOPJb0naaGktZJ2Vm/bKenhbjUJoHPX9Jvd9mJJ35Z0SNJQRIxVpTOShlrMMyxpuH6LAJow7b3xtudK+r2kH0bE+cm1mBgdcspBGyNie0Qsj4jlHXUKoCPTCrvtr2gi6L+JiD9Uk8/aXlDVF0ga706LAJrQdjPetiXtkPReRPxsUmmPpE2Snqked3elQ1y37rnnnmJ92bJlxfqRI0eabOe6N53f7PdJ+mdJx2xf+Xa3aSLkv7P9qKRTksoHTQH0VduwR8RBSVMO7i7pgWbbAdAtnC4LJEHYgSQIO5AEYQeSIOxAElziir5Zvrx8UuXWreVrqzZs2NBkO9c91uxAEoQdSIKwA0kQdiAJwg4kQdiBJAg7kATH2VF07NixfreAhrBmB5Ig7EAShB1IgrADSRB2IAnCDiRB2IEkOM6Ootdee61Yf/bZZ7u27MOHD3ftszNizQ4kQdiBJAg7kARhB5Ig7EAShB1IgrADSTgiym+wF0naJWlIUkjaHhG/sP2kpMck/W/11m0R8cc2n1VeGAbOrFnl9cGSJUuK9c2bN7esnThxojjvrl27ivVLly4V61lFxJSjLk/npJqLkn4UEW/a/pqkw7b3VrWfR8RPm2oSQPdMZ3z2MUlj1fOPbb8naWG3GwPQrGv6zW57saRvSzpUTXrc9lHbL9qe12KeYdsjtkc66hRAR6YddttzJf1e0g8j4rykX0r6hqRlmljzT3mSdERsj4jlEVEe2AtAV00r7La/oomg/yYi/iBJEXE2Ii5FxGVJv5K0onttAuhU27DbtqQdkt6LiJ9Nmr5g0tvWSTrefHsAmjKdQ28rJf2PpGOSLleTt0naqIlN+JB0UtKWamde6bM49AZ0WatDb23D3iTCDnRfq7BzBh2QBGEHkiDsQBKEHUiCsANJEHYgCcIOJEHYgSQIO5AEYQeSIOxAEoQdSIKwA0kQdiCJXg/Z/KGkU5Ne31JNG0SD2tug9iXRW11N9vaPrQo9vZ79Swu3Rwb13nSD2tug9iXRW1296o3NeCAJwg4k0e+wb+/z8ksGtbdB7Uuit7p60ltff7MD6J1+r9kB9AhhB5LoS9htP2j7z7Y/sL21Hz20Yvuk7WO2j/R7fLpqDL1x28cnTZtve6/t96vHKcfY61NvT9oerb67I7bX9Km3Rbb3237X9ju2f1BN7+t3V+irJ99bz3+z275B0l8krZJ0WtIbkjZGxLs9baQF2yclLY+Ivp+AYfs7ki5I2hUR36qm/ZukcxHxTPUP5byI+PGA9PakpAv9Hsa7Gq1oweRhxiU9LGmz+vjdFfparx58b/1Ys6+Q9EFEnIiIv0v6raS1fehj4EXEAUnnrpq8VtLO6vlOTfzP0nMtehsIETEWEW9Wzz+WdGWY8b5+d4W+eqIfYV8o6W+TXp/WYI33HpL+ZPuw7eF+NzOFoUnDbJ2RNNTPZqbQdhjvXrpqmPGB+e7qDH/eKXbQfdnKiPgnSQ9J+n61uTqQYuI32CAdO53WMN69MsUw45/r53dXd/jzTvUj7KOSFk16/fVq2kCIiNHqcVzSqxq8oajPXhlBt3oc73M/nxukYbynGmZcA/Dd9XP4836E/Q1Jd9heYvurkjZI2tOHPr7E9pxqx4lsz5G0WoM3FPUeSZuq55sk7e5jL18wKMN4txpmXH3+7vo+/HlE9PxP0hpN7JH/q6R/7UcPLfq6TdLb1d87/e5N0kua2Kz7P03s23hU0s2S9kl6X9J/S5o/QL39uyaG9j6qiWAt6FNvKzWxiX5U0pHqb02/v7tCXz353jhdFkiCHXRAEoQdSIKwA0kQdiAJwg4kQdiBJAg7kMT/A3x/IVSwjhIBAAAAAElFTkSuQmCC\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "YmwhWiPnRGoP",
        "colab_type": "code",
        "outputId": "5c6c40cb-6ee0-4b23-bef1-557745ef2c70",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 301
        }
      },
      "source": [
        "print(Y_train[index+1])\n",
        "plt.imshow(X_train[index+1].reshape((28,28)), cmap='gray')"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "7\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.image.AxesImage at 0x7f8a45474be0>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 7
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "rZjbgNKCLDaI",
        "colab_type": "code",
        "outputId": "4c289448-b2b1-44ed-d08f-aa4528d75406",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 88
        }
      },
      "source": [
        "X_train = X_train/255.0\n",
        "x_test = x_test/255.0\n",
        "final_Xtrain = torch.from_numpy(X_train.reshape(-1,1,28,28)).to(device)\n",
        "final_Ytrain = torch.from_numpy(Y_train.squeeze()).to(device)\n",
        "\n",
        "final_Xtest = torch.from_numpy(x_test.reshape(-1,1,28,28)).to(device)\n",
        "final_Ytest = torch.from_numpy(y_test.squeeze()).to(device)\n",
        "\n",
        "print(final_Xtrain.size())\n",
        "print(final_Ytrain.size())\n",
        "print(final_Xtest.size())\n",
        "print(final_Ytest.size())"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "torch.Size([175000, 1, 28, 28])\n",
            "torch.Size([175000])\n",
            "torch.Size([7000, 1, 28, 28])\n",
            "torch.Size([7000])\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "fvTmuMGLLJGN",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "train_ds = TensorDataset(final_Xtrain,final_Ytrain)\n",
        "test_ds = TensorDataset(final_Xtest,final_Ytest)\n",
        "\n",
        "train_dl = DataLoader(train_ds,batch_size=64,shuffle=True)\n",
        "test_dl = DataLoader(test_ds,batch_size=64,shuffle=True)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "mcV13UUfLMY7",
        "colab_type": "code",
        "outputId": "122f63de-102c-4b33-e4f8-3adcec656de8",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 355
        }
      },
      "source": [
        "class MNIST_CNN(nn.Module):\n",
        "  def __init__(self):\n",
        "    super(MNIST_CNN,self).__init__()\n",
        "    self.layer1 = nn.Sequential(\n",
        "            nn.Conv2d(1, 32, kernel_size=5, stride=1, padding=2),\n",
        "            nn.ReLU(),\n",
        "            nn.BatchNorm2d(32),\n",
        "            nn.MaxPool2d(2,2))\n",
        "    self.layer2 = nn.Sequential(\n",
        "            nn.Conv2d(32, 64, kernel_size=5, stride=1, padding=2),\n",
        "            nn.ReLU(),\n",
        "            nn.BatchNorm2d(64),\n",
        "            nn.MaxPool2d(2,2))\n",
        "    self.drop1 = nn.Dropout2d(0.4)\n",
        "    self.lin1 = nn.Linear(7*7*64,128)\n",
        "    self.norm = nn.BatchNorm1d(128)\n",
        "    self.drop2 = nn.Dropout(0.4)\n",
        "    self.lin2 = nn.Linear(128,10)\n",
        "\n",
        "\n",
        "\n",
        "  def forward(self,xb):\n",
        "    xb=xb.view(-1,1,28,28)\n",
        "    xb = self.layer1(xb)\n",
        "    xb = self.layer2(xb)\n",
        "    xb = self.drop1(xb)\n",
        "    xb = torch.flatten(xb,1,3)\n",
        "    xb = F.relu(self.lin1(xb))\n",
        "    xb = self.norm(xb)\n",
        "    xb = self.drop2(xb)\n",
        "    xb = self.lin2(xb)\n",
        "    return xb\n",
        "\n",
        "model = MNIST_CNN()\n",
        "model.float()\n",
        "model.to(device)      "
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "MNIST_CNN(\n",
              "  (layer1): Sequential(\n",
              "    (0): Conv2d(1, 32, kernel_size=(5, 5), stride=(1, 1), padding=(2, 2))\n",
              "    (1): ReLU()\n",
              "    (2): BatchNorm2d(32, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "    (3): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)\n",
              "  )\n",
              "  (layer2): Sequential(\n",
              "    (0): Conv2d(32, 64, kernel_size=(5, 5), stride=(1, 1), padding=(2, 2))\n",
              "    (1): ReLU()\n",
              "    (2): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "    (3): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)\n",
              "  )\n",
              "  (drop1): Dropout2d(p=0.4, inplace=False)\n",
              "  (lin1): Linear(in_features=3136, out_features=128, bias=True)\n",
              "  (norm): BatchNorm1d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "  (drop2): Dropout(p=0.4, inplace=False)\n",
              "  (lin2): Linear(in_features=128, out_features=10, bias=True)\n",
              ")"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 10
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "nBTbNR5OLQpJ",
        "colab_type": "code",
        "outputId": "97a641ac-ee33-416f-9b35-427a34e81a17",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        }
      },
      "source": [
        "criterion = nn.CrossEntropyLoss()\n",
        "optimizer = optim.Adam(model.parameters(),lr=0.001)\n",
        "lr_scheduler = torch.optim.lr_scheduler.ReduceLROnPlateau(optimizer, patience=1, verbose=True, eps=1e-09)\n",
        "\n",
        "epochs = 50\n",
        "costs = []\n",
        "\n",
        "for epoch in range(epochs):\n",
        "  total_loss = 0\n",
        "  train_acc = 0\n",
        "  model.train()\n",
        "  for (x,y) in train_dl:\n",
        "    x.to(device)\n",
        "    y.to(device)\n",
        "    Ypred = model(x.float())\n",
        "    Ypred.to(device)\n",
        "    loss = criterion(Ypred,y)\n",
        "    total_loss = total_loss + loss\n",
        "    Ypred_binary = torch.argmax(Ypred, keepdim= False, dim=1)\n",
        "    train_acc = train_acc + accuracy_score(y.cpu(),Ypred_binary.cpu())\n",
        "    optimizer.zero_grad()\n",
        "    loss.backward()\n",
        "    optimizer.step()\n",
        "\n",
        "  costs.append(total_loss)\n",
        "  print(\"Epoch No. = \"+ str(epoch+1), \"cost = \"+str(costs[epoch].item()))\n",
        "  print(\"Training Accuracy = \", train_acc/len(train_dl))\n",
        "  lr_scheduler.step(total_loss/len(train_dl))\n",
        "\n",
        "  model.eval()\n",
        "  with torch.no_grad():\n",
        "    cv_acc = 0\n",
        "    for x_val,y_val in test_dl:\n",
        "      x_val.to(device)\n",
        "      y_val.to(device)\n",
        "      Ypred_val = model(x_val.float())\n",
        "      Ypred_val.to(device)\n",
        "      Ypred_val_binary = torch.argmax(Ypred_val, keepdim= False, dim=1)\n",
        "      cv_acc = cv_acc + accuracy_score(y_val.cpu(),Ypred_val_binary.cpu())\n",
        "    print(\"Test Accuracy = \", cv_acc/len(test_dl))\n",
        "\n",
        "plt.plot(costs)\n",
        "plt.ylabel(\"cost\")\n",
        "plt.xlabel(\"epoch\")\n",
        "plt.show()      "
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Epoch No. = 1 cost = 267.1976013183594\n",
            "Training Accuracy =  0.9713913010359537\n",
            "Test Accuracy =  0.9861268939393939\n",
            "Epoch No. = 2 cost = 122.56755065917969\n",
            "Training Accuracy =  0.9858051492992078\n",
            "Test Accuracy =  0.9899147727272727\n",
            "Epoch No. = 3 cost = 95.4149169921875\n",
            "Training Accuracy =  0.9888368372943327\n",
            "Test Accuracy =  0.9910511363636364\n",
            "Epoch No. = 4 cost = 78.2820816040039\n",
            "Training Accuracy =  0.9907221206581353\n",
            "Test Accuracy =  0.9914772727272727\n",
            "Epoch No. = 5 cost = 64.26144409179688\n",
            "Training Accuracy =  0.9925293266301036\n",
            "Test Accuracy =  0.9914772727272727\n",
            "Epoch No. = 6 cost = 57.17283248901367\n",
            "Training Accuracy =  0.993289152955515\n",
            "Test Accuracy =  0.9919034090909091\n",
            "Epoch No. = 7 cost = 50.57075500488281\n",
            "Training Accuracy =  0.9943041590493601\n",
            "Test Accuracy =  0.9914772727272727\n",
            "Epoch No. = 8 cost = 44.61053466796875\n",
            "Training Accuracy =  0.9946755027422304\n",
            "Test Accuracy =  0.990909090909091\n",
            "Epoch No. = 9 cost = 41.945438385009766\n",
            "Training Accuracy =  0.9948811700182816\n",
            "Test Accuracy =  0.9916193181818181\n",
            "Epoch No. = 10 cost = 39.541847229003906\n",
            "Training Accuracy =  0.9951725319926874\n",
            "Test Accuracy =  0.9920454545454546\n",
            "Epoch No. = 11 cost = 37.05317306518555\n",
            "Training Accuracy =  0.9955095978062157\n",
            "Test Accuracy =  0.9931818181818182\n",
            "Epoch No. = 12 cost = 34.12224197387695\n",
            "Training Accuracy =  0.9959152193784278\n",
            "Test Accuracy =  0.9923295454545454\n",
            "Epoch No. = 13 cost = 30.580875396728516\n",
            "Training Accuracy =  0.9962751371115174\n",
            "Test Accuracy =  0.9910984848484848\n",
            "Epoch No. = 14 cost = 28.797733306884766\n",
            "Training Accuracy =  0.9964541438147472\n",
            "Test Accuracy =  0.9926136363636363\n",
            "Epoch No. = 15 cost = 26.920711517333984\n",
            "Training Accuracy =  0.9967664533820841\n",
            "Test Accuracy =  0.9923295454545454\n",
            "Epoch No. = 16 cost = 24.3087100982666\n",
            "Training Accuracy =  0.9970063985374772\n",
            "Test Accuracy =  0.9916193181818181\n",
            "Epoch No. = 17 cost = 28.07640838623047\n",
            "Training Accuracy =  0.9965398385131018\n",
            "Test Accuracy =  0.9928977272727273\n",
            "Epoch No. = 18 cost = 24.012893676757812\n",
            "Training Accuracy =  0.9971092321755027\n",
            "Test Accuracy =  0.9923295454545454\n",
            "Epoch No. = 19 cost = 22.217018127441406\n",
            "Training Accuracy =  0.9972806215722121\n",
            "Test Accuracy =  0.9919034090909091\n",
            "Epoch No. = 20 cost = 22.803260803222656\n",
            "Training Accuracy =  0.9971777879341864\n",
            "Test Accuracy =  0.9927556818181819\n",
            "Epoch No. = 21 cost = 21.756168365478516\n",
            "Training Accuracy =  0.9972234917733089\n",
            "Test Accuracy =  0.9919507575757576\n",
            "Epoch No. = 22 cost = 20.41349983215332\n",
            "Training Accuracy =  0.9975377056672761\n",
            "Test Accuracy =  0.9914772727272727\n",
            "Epoch No. = 23 cost = 20.318092346191406\n",
            "Training Accuracy =  0.9974691499085924\n",
            "Test Accuracy =  0.9921875\n",
            "Epoch No. = 24 cost = 19.242036819458008\n",
            "Training Accuracy =  0.9977719378427788\n",
            "Test Accuracy =  0.992092803030303\n",
            "Epoch No. = 25 cost = 18.007585525512695\n",
            "Training Accuracy =  0.9978005027422303\n",
            "Test Accuracy =  0.9923295454545454\n",
            "Epoch No. = 26 cost = 17.8132266998291\n",
            "Training Accuracy =  0.9977605118829982\n",
            "Test Accuracy =  0.9926136363636363\n",
            "Epoch No. = 27 cost = 17.594297409057617\n",
            "Training Accuracy =  0.9977776508226691\n",
            "Test Accuracy =  0.9917613636363637\n",
            "Epoch No. = 28 cost = 17.492298126220703\n",
            "Training Accuracy =  0.9978176416819012\n",
            "Test Accuracy =  0.9923295454545454\n",
            "Epoch No. = 29 cost = 16.483789443969727\n",
            "Training Accuracy =  0.9980575868372943\n",
            "Test Accuracy =  0.9920454545454546\n",
            "Epoch No. = 30 cost = 16.088367462158203\n",
            "Training Accuracy =  0.9980175959780622\n",
            "Test Accuracy =  0.9923295454545454\n",
            "Epoch No. = 31 cost = 15.66380500793457\n",
            "Training Accuracy =  0.9981489945155393\n",
            "Test Accuracy =  0.9921875\n",
            "Epoch No. = 32 cost = 15.866164207458496\n",
            "Training Accuracy =  0.9980118829981719\n",
            "Test Accuracy =  0.9928977272727273\n",
            "Epoch No. = 33 cost = 15.532546997070312\n",
            "Training Accuracy =  0.9980804387568556\n",
            "Test Accuracy =  0.9924715909090909\n",
            "Epoch No. = 34 cost = 14.223441123962402\n",
            "Training Accuracy =  0.9981718464351006\n",
            "Test Accuracy =  0.9912405303030303\n",
            "Epoch No. = 35 cost = 13.447915077209473\n",
            "Training Accuracy =  0.9983432358318098\n",
            "Test Accuracy =  0.99375\n",
            "Epoch No. = 36 cost = 15.652904510498047\n",
            "Training Accuracy =  0.9980975776965265\n",
            "Test Accuracy =  0.9924715909090909\n",
            "Epoch No. = 37 cost = 14.848665237426758\n",
            "Training Accuracy =  0.9982575411334552\n",
            "Epoch    37: reducing learning rate of group 0 to 1.0000e-04.\n",
            "Test Accuracy =  0.9923295454545454\n",
            "Epoch No. = 38 cost = 10.671808242797852\n",
            "Training Accuracy =  0.9986745886654479\n",
            "Test Accuracy =  0.9921875\n",
            "Epoch No. = 39 cost = 8.72739315032959\n",
            "Training Accuracy =  0.9989373857404021\n",
            "Test Accuracy =  0.9922348484848484\n",
            "Epoch No. = 40 cost = 7.356316566467285\n",
            "Training Accuracy =  0.9991659049360146\n",
            "Test Accuracy =  0.9928977272727273\n",
            "Epoch No. = 41 cost = 7.150975227355957\n",
            "Training Accuracy =  0.9991202010968921\n",
            "Test Accuracy =  0.9931818181818182\n",
            "Epoch No. = 42 cost = 6.130136013031006\n",
            "Training Accuracy =  0.9992515996343693\n",
            "Test Accuracy =  0.9925662878787879\n",
            "Epoch No. = 43 cost = 5.5948591232299805\n",
            "Training Accuracy =  0.9993258683729433\n",
            "Test Accuracy =  0.9931818181818182\n",
            "Epoch No. = 44 cost = 5.183516025543213\n",
            "Training Accuracy =  0.9993829981718464\n",
            "Test Accuracy =  0.9928030303030303\n",
            "Epoch No. = 45 cost = 4.828579425811768\n",
            "Training Accuracy =  0.999337294332724\n",
            "Test Accuracy =  0.9927556818181819\n",
            "Epoch No. = 46 cost = 5.146921634674072\n",
            "Training Accuracy =  0.9993487202925045\n",
            "Test Accuracy =  0.9929450757575757\n",
            "Epoch No. = 47 cost = 5.250174045562744\n",
            "Training Accuracy =  0.9994020414381475\n",
            "Epoch    47: reducing learning rate of group 0 to 1.0000e-05.\n",
            "Test Accuracy =  0.9928977272727273\n",
            "Epoch No. = 48 cost = 4.664043426513672\n",
            "Training Accuracy =  0.9993258683729433\n",
            "Test Accuracy =  0.9930397727272727\n",
            "Epoch No. = 49 cost = 4.72037935256958\n",
            "Training Accuracy =  0.9994686928702011\n",
            "Test Accuracy =  0.9928977272727273\n",
            "Epoch No. = 50 cost = 3.76944899559021\n",
            "Training Accuracy =  0.9995372486288848\n",
            "Test Accuracy =  0.9933238636363636\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "xeXutUyKMD1Y",
        "colab_type": "code",
        "outputId": "c3d63ea6-0b59-4acd-df6c-c1b66f1d6177",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 372
        }
      },
      "source": [
        "epsilons = [0, .05, .1, .15, .2, .25, .3]\n",
        "print(\"CUDA Available: \",torch.cuda.is_available())\n",
        "device = torch.device(\"cuda\" if (torch.cuda.is_available()) else \"cpu\")\n",
        "test_loader = DataLoader(test_ds,batch_size=1,shuffle=True)\n",
        "model.eval()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "CUDA Available:  True\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "MNIST_CNN(\n",
              "  (layer1): Sequential(\n",
              "    (0): Conv2d(1, 32, kernel_size=(5, 5), stride=(1, 1), padding=(2, 2))\n",
              "    (1): ReLU()\n",
              "    (2): BatchNorm2d(32, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "    (3): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)\n",
              "  )\n",
              "  (layer2): Sequential(\n",
              "    (0): Conv2d(32, 64, kernel_size=(5, 5), stride=(1, 1), padding=(2, 2))\n",
              "    (1): ReLU()\n",
              "    (2): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "    (3): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)\n",
              "  )\n",
              "  (drop1): Dropout2d(p=0.4, inplace=False)\n",
              "  (lin1): Linear(in_features=3136, out_features=128, bias=True)\n",
              "  (norm): BatchNorm1d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "  (drop2): Dropout(p=0.4, inplace=False)\n",
              "  (lin2): Linear(in_features=128, out_features=10, bias=True)\n",
              ")"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 12
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "-BybNpLTMJJa",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def fgsm_attack(image, epsilon, data_grad):\n",
        "    sign_data_grad = data_grad.sign()\n",
        "    perturbed_image = image + epsilon*sign_data_grad\n",
        "    perturbed_image = torch.clamp(perturbed_image, 0, 1)\n",
        "    return perturbed_image"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "1ZcDA6pfMf13",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def test( model, device, test_loader, epsilon ):\n",
        "\n",
        "    correct = 0\n",
        "    adv_examples = []\n",
        "\n",
        "    for data, target in test_loader:\n",
        "\n",
        "        data, target = data.to(device), target.to(device)\n",
        "\n",
        "        data.requires_grad = True\n",
        "\n",
        "        output = model(data)\n",
        "        init_pred = output.max(1, keepdim=True)[1] \n",
        "\n",
        "        if init_pred.item() != target.item():\n",
        "            continue\n",
        "\n",
        "        loss = F.nll_loss(output, target)\n",
        "\n",
        "        model.zero_grad()\n",
        "\n",
        "        loss.backward()\n",
        "\n",
        "        data_grad = data.grad.data\n",
        "\n",
        "        perturbed_data = fgsm_attack(data, epsilon, data_grad)\n",
        "\n",
        "        output = model(perturbed_data)\n",
        "\n",
        "        final_pred = output.max(1, keepdim=True)[1] \n",
        "        if final_pred.item() == target.item():\n",
        "            correct += 1\n",
        "            if (epsilon == 0) and (len(adv_examples) < 5):\n",
        "                adv_ex = perturbed_data.squeeze().detach().cpu().numpy()\n",
        "                adv_examples.append( (init_pred.item(), final_pred.item(), adv_ex) )\n",
        "        else:\n",
        "            if len(adv_examples) < 5:\n",
        "                adv_ex = perturbed_data.squeeze().detach().cpu().numpy()\n",
        "                adv_examples.append( (init_pred.item(), final_pred.item(), adv_ex) )\n",
        "\n",
        "    final_acc = correct/float(len(test_loader))\n",
        "    print(\"Epsilon: {}\\tTest Accuracy = {} / {} = {}\".format(epsilon, correct, len(test_loader), final_acc))\n",
        "\n",
        "    return final_acc, adv_examples"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "LfTYQYz7M_x4",
        "colab_type": "code",
        "outputId": "1bacb663-5d51-4afb-ee88-b3ad2a57e056",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 141
        }
      },
      "source": [
        "accuracies = []\n",
        "examples = []\n",
        "\n",
        "# Run test for each epsilon\n",
        "for eps in epsilons:\n",
        "    acc, ex = test(model, device, test_loader, eps)\n",
        "    accuracies.append(acc)\n",
        "    examples.append(ex)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Epsilon: 0\tTest Accuracy = 6953 / 7000 = 0.9932857142857143\n",
            "Epsilon: 0.05\tTest Accuracy = 6453 / 7000 = 0.9218571428571428\n",
            "Epsilon: 0.1\tTest Accuracy = 4588 / 7000 = 0.6554285714285715\n",
            "Epsilon: 0.15\tTest Accuracy = 2413 / 7000 = 0.3447142857142857\n",
            "Epsilon: 0.2\tTest Accuracy = 1136 / 7000 = 0.16228571428571428\n",
            "Epsilon: 0.25\tTest Accuracy = 669 / 7000 = 0.09557142857142857\n",
            "Epsilon: 0.3\tTest Accuracy = 475 / 7000 = 0.06785714285714285\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "DCYz6PhUNgA3",
        "colab_type": "code",
        "outputId": "c04a9a58-bac2-4cf0-ccb6-1b815a797af1",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 350
        }
      },
      "source": [
        "plt.figure(figsize=(5,5))\n",
        "plt.plot(epsilons, accuracies, \"*-\")\n",
        "plt.yticks(np.arange(0, 1.1, step=0.1))\n",
        "plt.xticks(np.arange(0, .35, step=0.05))\n",
        "plt.title(\"Accuracy vs Epsilon\")\n",
        "plt.xlabel(\"Epsilon\")\n",
        "plt.ylabel(\"Accuracy\")\n",
        "plt.show()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 360x360 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "UX6dvWaTNlXp",
        "colab_type": "code",
        "outputId": "86378966-d26a-4853-ada7-909b729a7bdf",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 729
        }
      },
      "source": [
        "cnt = 0\n",
        "plt.figure(figsize=(8,10))\n",
        "for i in range(len(epsilons)):\n",
        "    for j in range(len(examples[i])):\n",
        "        cnt += 1\n",
        "        plt.subplot(len(epsilons),len(examples[0]),cnt)\n",
        "        plt.xticks([], [])\n",
        "        plt.yticks([], [])\n",
        "        if j == 0:\n",
        "            plt.ylabel(\"Eps: {}\".format(epsilons[i]), fontsize=14)\n",
        "        orig,adv,ex = examples[i][j]\n",
        "        plt.title(\"{} -> {}\".format(orig, adv))\n",
        "        plt.imshow(ex, cmap=\"gray\")\n",
        "plt.tight_layout()\n",
        "plt.show()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 576x720 with 35 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "RxnagpDZ2XFZ",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "x_train = x_train/255.0\n",
        "xtrain = torch.from_numpy(x_train.reshape(-1,1,28,28)).to(device)\n",
        "ytrain = torch.from_numpy(y_train.squeeze()).to(device)\n",
        "original_ds = TensorDataset(xtrain,ytrain)\n",
        "original_dl = DataLoader(original_ds,batch_size=1,shuffle=False)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "_RiWEicP3wUJ",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def perturbed_images(model, device, data_loader, epsilon):\n",
        "\n",
        "  perturbed_data = []\n",
        "\n",
        "  for data, target in data_loader:\n",
        "\n",
        "    data, target = data.to(device), target.to(device)\n",
        "    data.requires_grad = True\n",
        "    \n",
        "    output = model(data)\n",
        "    loss = F.nll_loss(output, target)\n",
        "    model.zero_grad()\n",
        "    loss.backward()\n",
        "    data_grad = data.grad.data\n",
        "    perturbed_image = fgsm_attack(data, epsilon, data_grad)\n",
        "    perturbed_ex = perturbed_image.squeeze().detach().cpu().numpy()\n",
        "    perturbed_data.append(perturbed_ex)\n",
        "\n",
        "  return perturbed_data  "
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "JOFre7z-9u1U",
        "colab_type": "code",
        "outputId": "28276076-b148-42b8-a742-9492340330cc",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 52
        }
      },
      "source": [
        "perturbed_eps1 = perturbed_images(model, device, original_dl, 0.05)\n",
        "print(len(perturbed_eps1))\n",
        "print(perturbed_eps1[0].shape)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "35000\n",
            "(28, 28)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "a5BkpP3LMuyX",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "for i in range(len(perturbed_eps1)):\n",
        "  perturbed_eps1[i] = perturbed_eps1[i].reshape((784))"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "k2vTbQmDO0uw",
        "colab_type": "code",
        "outputId": "72112f78-aabd-4718-9ad6-efe7b36da200",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 70
        }
      },
      "source": [
        "perturbed_eps2 = perturbed_images(model, device, original_dl, 0.1)\n",
        "for i in range(len(perturbed_eps2)):\n",
        "  perturbed_eps2[i] = perturbed_eps2[i].reshape((784))\n",
        "\n",
        "print(perturbed_eps1[0].shape)\n",
        "print(perturbed_eps2[1].shape)\n",
        "print(len(perturbed_eps2)) "
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "(784,)\n",
            "(784,)\n",
            "35000\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "2oSKwY9nP6al",
        "colab_type": "code",
        "outputId": "cb23cb67-6245-49bd-d2bf-af7dde24eb02",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 52
        }
      },
      "source": [
        "perturbed_eps1 = np.asarray(perturbed_eps1,dtype=np.float32)\n",
        "perturbed_eps2 = np.asarray(perturbed_eps2,dtype=np.float32)\n",
        "print(perturbed_eps1.shape)\n",
        "print(perturbed_eps2.shape)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "(35000, 784)\n",
            "(35000, 784)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Lv3_1evIQWBl",
        "colab_type": "code",
        "outputId": "c845e0cc-a4cf-4fc8-998a-2a36356b471a",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 52
        }
      },
      "source": [
        "new_Xtrain = np.concatenate((perturbed_eps1,perturbed_eps2),axis=0)\n",
        "new_Ytrain = np.concatenate((y_train,y_train),axis=0)\n",
        "print(new_Xtrain.shape)\n",
        "print(new_Ytrain.shape)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "(70000, 784)\n",
            "(70000,)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "PfPFKEs2RTd6",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "new_xtrain = torch.from_numpy(new_Xtrain.reshape(-1,1,28,28)).to(device)\n",
        "new_ytrain = torch.from_numpy(new_Ytrain.squeeze()).to(device)\n",
        "new_train_ds = TensorDataset(new_xtrain,new_ytrain)\n",
        "train_loader = DataLoader(new_train_ds,batch_size=64,shuffle=True)\n",
        "#new_model = MNIST_CNN()\n",
        "#new_model.float()\n",
        "#new_model.to(device)      "
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Nwy4-fq4V61A",
        "colab_type": "code",
        "outputId": "b3f65a99-1a49-40a0-8883-0678122ef7aa",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 301
        }
      },
      "source": [
        "print(new_Ytrain[1])\n",
        "plt.imshow(new_Xtrain[1].reshape((28,28)), cmap='gray')"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "7\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.image.AxesImage at 0x7f8a4e4a2278>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 26
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "4YsNlCesWc7s",
        "colab_type": "code",
        "outputId": "298385d9-b259-45bc-81a8-78a3b5e878ae",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 301
        }
      },
      "source": [
        "print(new_Ytrain[1+35000])\n",
        "plt.imshow(new_Xtrain[1+35000].reshape((28,28)), cmap='gray')"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "7\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.image.AxesImage at 0x7f8a6011ceb8>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 27
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "SBDnsZcHSezc",
        "colab_type": "code",
        "outputId": "367249bf-dcff-42ef-acb9-dbdd53c90378",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        }
      },
      "source": [
        "#criterion = nn.CrossEntropyLoss()\n",
        "#optimizer = optim.Adam(model.parameters(),lr=0.001)\n",
        "#lr_scheduler = torch.optim.lr_scheduler.ReduceLROnPlateau(optimizer, patience=1, verbose=True, eps=1e-09)\n",
        "\n",
        "epochs = 20\n",
        "costs = []\n",
        "\n",
        "for epoch in range(epochs):\n",
        "  total_loss = 0\n",
        "  train_acc = 0\n",
        "  model.train()\n",
        "  for (x,y) in train_loader:\n",
        "    x.to(device)\n",
        "    y.to(device)\n",
        "    Ypred = model(x.float())\n",
        "    Ypred.to(device)\n",
        "    loss = criterion(Ypred,y)\n",
        "    total_loss = total_loss + loss\n",
        "    Ypred_binary = torch.argmax(Ypred, keepdim= False, dim=1)\n",
        "    train_acc = train_acc + accuracy_score(y.cpu(),Ypred_binary.cpu())\n",
        "    optimizer.zero_grad()\n",
        "    loss.backward()\n",
        "    optimizer.step()\n",
        "\n",
        "  costs.append(total_loss)\n",
        "  print(\"Epoch No. = \"+ str(epoch+1), \"cost = \"+str(costs[epoch].item()))\n",
        "  print(\"Training Accuracy = \", train_acc/len(train_loader))\n",
        "  lr_scheduler.step(total_loss/len(train_loader))\n",
        "\n",
        "  model.eval()\n",
        "  with torch.no_grad():\n",
        "    cv_acc = 0\n",
        "    for x_val,y_val in test_dl:\n",
        "      x_val.to(device)\n",
        "      y_val.to(device)\n",
        "      Ypred_val = model(x_val.float())\n",
        "      Ypred_val.to(device)\n",
        "      Ypred_val_binary = torch.argmax(Ypred_val, keepdim= False, dim=1)\n",
        "      cv_acc = cv_acc + accuracy_score(y_val.cpu(),Ypred_val_binary.cpu())\n",
        "    print(\"Test Accuracy = \", cv_acc/len(test_dl))\n",
        "\n",
        "plt.plot(costs)\n",
        "plt.ylabel(\"cost\")\n",
        "plt.xlabel(\"epoch\")\n",
        "plt.show()      "
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Epoch No. = 1 cost = 887.2261352539062\n",
            "Training Accuracy =  0.8793275822669104\n",
            "Test Accuracy =  0.9916193181818181\n",
            "Epoch No. = 2 cost = 648.0170288085938\n",
            "Training Accuracy =  0.9045837141986593\n",
            "Epoch    52: reducing learning rate of group 0 to 1.0000e-06.\n",
            "Test Accuracy =  0.9919034090909091\n",
            "Epoch No. = 3 cost = 572.5530395507812\n",
            "Training Accuracy =  0.9122771937842779\n",
            "Test Accuracy =  0.9914772727272727\n",
            "Epoch No. = 4 cost = 557.6246337890625\n",
            "Training Accuracy =  0.9146528412553322\n",
            "Epoch    54: reducing learning rate of group 0 to 1.0000e-07.\n",
            "Test Accuracy =  0.9911931818181818\n",
            "Epoch No. = 5 cost = 549.8146362304688\n",
            "Training Accuracy =  0.9147909049360147\n",
            "Test Accuracy =  0.9914772727272727\n",
            "Epoch No. = 6 cost = 551.3003540039062\n",
            "Training Accuracy =  0.9138625457038391\n",
            "Epoch    56: reducing learning rate of group 0 to 1.0000e-08.\n",
            "Test Accuracy =  0.9917613636363637\n",
            "Epoch No. = 7 cost = 560.9049682617188\n",
            "Training Accuracy =  0.9151860527117611\n",
            "Test Accuracy =  0.9916193181818181\n",
            "Epoch No. = 8 cost = 553.1797485351562\n",
            "Training Accuracy =  0.915366963741621\n",
            "Epoch    58: reducing learning rate of group 0 to 1.0000e-09.\n",
            "Test Accuracy =  0.9917613636363637\n",
            "Epoch No. = 9 cost = 554.5791015625\n",
            "Training Accuracy =  0.9149289686166971\n",
            "Test Accuracy =  0.9914772727272727\n",
            "Epoch No. = 10 cost = 541.736572265625\n",
            "Training Accuracy =  0.9157621115173674\n",
            "Test Accuracy =  0.9917613636363637\n",
            "Epoch No. = 11 cost = 556.4910278320312\n",
            "Training Accuracy =  0.9147290143205362\n",
            "Test Accuracy =  0.9916193181818181\n",
            "Epoch No. = 12 cost = 556.4612426757812\n",
            "Training Accuracy =  0.9151622486288848\n",
            "Test Accuracy =  0.9916193181818181\n",
            "Epoch No. = 13 cost = 547.96533203125\n",
            "Training Accuracy =  0.9168666209628276\n",
            "Test Accuracy =  0.9916193181818181\n",
            "Epoch No. = 14 cost = 545.6323852539062\n",
            "Training Accuracy =  0.9165571678854357\n",
            "Test Accuracy =  0.9914772727272727\n",
            "Epoch No. = 15 cost = 553.0548706054688\n",
            "Training Accuracy =  0.914448126142596\n",
            "Test Accuracy =  0.9912405303030303\n",
            "Epoch No. = 16 cost = 547.2102661132812\n",
            "Training Accuracy =  0.916471473187081\n",
            "Test Accuracy =  0.9916193181818181\n",
            "Epoch No. = 17 cost = 560.033447265625\n",
            "Training Accuracy =  0.9142767367458866\n",
            "Test Accuracy =  0.9916193181818181\n",
            "Epoch No. = 18 cost = 550.5968627929688\n",
            "Training Accuracy =  0.9140672608165753\n",
            "Test Accuracy =  0.9916193181818181\n",
            "Epoch No. = 19 cost = 552.4263916015625\n",
            "Training Accuracy =  0.9147052102376599\n",
            "Test Accuracy =  0.9913825757575757\n",
            "Epoch No. = 20 cost = 551.4375\n",
            "Training Accuracy =  0.9153812461913468\n",
            "Test Accuracy =  0.9911931818181818\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYUAAAEICAYAAACwDehOAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjEsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+j8jraAAAgAElEQVR4nO3deXQc5Znv8e+jfbGtthbbWgze2W1ja7DZDMEkLAkxAZIhMxkI4Y7JDFlI7twJ3LkzYXJPzp1swx3uZEgghECGEBIDAyHsJGGNIcbYxhhjy7tkWZZlS5Yla3/uH12S20ISkqXqltW/zzl9uvqtqu5HrVb/VG/VW2XujoiICEBKogsQEZHRQ6EgIiI9FAoiItJDoSAiIj0UCiIi0kOhICIiPUINBTP7qpmtN7N3zeyWoC3fzJ43s83B/cSg3czsTjOrMLN1ZrYgzNpEROSDLKxxCmZ2OvBL4CygDXgG+CKwHNjv7v9iZrcCE939G2Z2OfBl4HJgEfBv7r5ooNcoLCz0adOmhVK/iMhY9dZbb+1z96K+5qWF+LqnAG+4ezOAmb0EXAUsAy4Mlrkf+APwjaD9AY+m1Eozi5hZsbtX9/cC06ZNY9WqVeH9BCIiY5CZ7ehvXpjdR+uB882swMxyiG4BTAUmx3zR7wEmB9OlwK6Y9SuDNhERiZPQthTc/T0z+w7wHNAErAE6ey3jZjak/iszW060C4oTTjhhhKoVEREIeUezu9/r7gvdfQlwANgE1JhZMUBwvzdYvIrolkS3sqCt93Pe7e7l7l5eVNRnl5iIiByjsI8+mhTcn0B0f8IvgCeA64NFrgceD6afAK4LjkJaDDQMtD9BRERGXpg7mgEeMbMCoB242d3rzexfgF+Z2Y3ADuAzwbJPEd3vUAE0AzeEXJuIiPQSaii4+/l9tNUBS/tod+DmMOsREZGBaUSziIj0SMpQ2LjnIN99ZiMNze2JLkVEZFRJylDYWdfMf/xhCzv3Nye6FBGRUSUpQ6Ekkg1AVf3hBFciIjK6JGUolAahsFuhICJylKQMhUhOOtnpqQoFEZFekjIUzIySSBa7GxQKIiKxkjIUILpfoaq+JdFliIiMKkkbCqWRbHUfiYj0krShUBLJpraxldaOzg9fWEQkSSR1KADsaVAXkohItyQOhSxAYxVERGIlbyjkdY9V0JaCiEi3pA2FKXnRLQXtbBYROSJpQyErPZXCcZkKBRGRGEkbCgClkSztUxARiZHUoVCisQoiIkdRKNS3EL3om4iIJH0oHG7vpF4X2xERAZI8FEo1VkFE5CihhoKZfc3M3jWz9Wb2kJllmdnPzGybma0JbvODZc3M7jSzCjNbZ2YLwqwNjoxq1n4FEZGotLCe2MxKga8Ap7r7YTP7FXBtMPt/uPuKXqtcBswObouAu4L70CgURESOFnb3URqQbWZpQA6we4BllwEPeNRKIGJmxWEWV5CbQUZaCrt1/iMRESDEUHD3KuD7wE6gGmhw9+eC2d8OuojuMLPMoK0U2BXzFJVBW2jMTKfQFhGJEVoomNlEov/9TwdKgFwz+xxwG3Ay8GdAPvCNIT7vcjNbZWaramtrh11nSSRLoSAiEgiz++hiYJu717p7O/AocI67VwddRK3AfcBZwfJVwNSY9cuCtqO4+93uXu7u5UVFRcMusiQvWyfFExEJhBkKO4HFZpZjZgYsBd7r3k8QtF0JrA+WfwK4LjgKaTHR7qbqEOsDojubaxpbaO/sCvulRERGvdCOPnL3N8xsBbAa6ADeBu4GnjazIsCANcAXg1WeAi4HKoBm4IawaotVGsnGPXqxnan5OfF4SRGRUSu0UABw928C3+zVfFE/yzpwc5j19CX2sFSFgogku6Qe0QxHrsC2u0E7m0VEFAoRXYFNRKRb0odCVnoqBbkZOv+RiAgKBUDXVRAR6aZQQAPYRES6KRSIbilUHTisi+2ISNJTKBAdq9DU1snBlo5ElyIiklAKBXQKbRGRbgoFFAoiIt0UCkBJXjCATaEgIklOoQAUjsskPdWo0gA2EUlyCgUgJcUoztNYBRERhUJAYxVERBQKPTSqWUREodCjNJLNnoMtdOhiOyKSxBQKgZJINl0ONY2tiS5FRCRhFAoBjVUQEVEo9CiNaKyCiIhCIVCcF91S0HUVRCSZhRoKZvY1M3vXzNab2UNmlmVm083sDTOrMLOHzSwjWDYzeFwRzJ8WZm295WamEclJ15aCiCS10ELBzEqBrwDl7n46kApcC3wHuMPdZwEHgBuDVW4EDgTtdwTLxVVJXrYuyykiSS3s7qM0INvM0oAcoBq4CFgRzL8fuDKYXhY8Jpi/1Mws5PqOorEKIpLsQgsFd68Cvg/sJBoGDcBbQL27d1+4oBIoDaZLgV3Buh3B8gVh1deX0kiW9imISFILs/toItH//qcDJUAucOkIPO9yM1tlZqtqa2uH+3RHKYlk09jSQWNL+4g+r4jI8SLM7qOLgW3uXuvu7cCjwLlAJOhOAigDqoLpKmAqQDA/D6jr/aTufre7l7t7eVFR0YgW3D1WobpB+xVEJDmFGQo7gcVmlhPsG1gKbAB+D1wTLHM98Hgw/UTwmGD+7zzOF03uDgV1IYlIsgpzn8IbRHcYrwbeCV7rbuAbwNfNrILoPoN7g1XuBQqC9q8Dt4ZVW39KNapZRJJc2ocvcuzc/ZvAN3s1bwXO6mPZFuDTYdbzYYrGZ5KWYgoFEUlaGtEcIzXFmJKXpbEKIpK0FAq9lESytU9BRJKWQqGXUg1gE5EkplDopSSSxZ6GFjq74nrgk4jIqKBQ6KUkkk1Hl1Ori+2ISBJSKPSisQoikswUCr1orIKIJDOFQi/FeboCm4gkL4VCL+Oz0hmflaZQEJGkpFDoQ2kkmyoNYBORJKRQ6IMutiMiyUqh0IeSSBa7GxQKIpJ8FAp9KIlkU9/cTlNrx4cvLCIyhigU+lDac7EdbS2ISHJRKPThyAA27WwWkeSiUOhDiQawiUiSUij0YfL4TFJMoSAiyUeh0Ie01BSmTMjS+Y9EJOkoFPqhsQoikowUCv2IhoJ2NItIcgktFMzsJDNbE3M7aGa3mNntZlYV0355zDq3mVmFmb1vZpeEVdtglESyqW44TJcutiMiSSQtrCd29/eB+QBmlgpUAY8BNwB3uPv3Y5c3s1OBa4HTgBLgBTOb4+6dYdU4kNJIFu2dzr5DrUyakJWIEkRE4i5e3UdLgS3uvmOAZZYBv3T3VnffBlQAZ8Wluj7oYjsikoziFQrXAg/FPP6Sma0zs5+a2cSgrRTYFbNMZdCWEEfGKmi/gogkj9BDwcwygE8Cvw6a7gJmEu1aqgZ+MMTnW25mq8xsVW1t7YjWGksD2EQkGcVjS+EyYLW71wC4e427d7p7F3APR7qIqoCpMeuVBW1Hcfe73b3c3cuLiopCK3pCVhrjMtPUfSQiSSUeofBZYrqOzKw4Zt6ngPXB9BPAtWaWaWbTgdnAm3Gor09mRkkkSyfFE5GkEtrRRwBmlgt8FLgppvm7ZjYfcGB79zx3f9fMfgVsADqAmxN15FE3jVUQkWQTaii4exNQ0KvtrwZY/tvAt8OsaShKItm8U9mQ6DJEROJGI5oHUBrJpq6pjZb2hG6wiIjEjUJhACWR6KA1HYEkIslCoTCAkjyNVRCR5KJQGIDGKohIslEoDGBKXhZmOtWFiCQPhcIA0lNTmDQ+U1sKIpI0FAofoiSSzW4NYBORJDGoUDCzTw+mbSzSADYRSSaD3VK4bZBtY05pJJuq+sO462I7IjL2DTii2cwuAy4HSs3szphZE4ieimLMK8nLoq2ji7qmNgrHZSa6HBGRUH3YaS52A6uInvr6rZj2RuBrYRU1msQelqpQEJGxbsBQcPe1wFoz+4W7twMEF8WZ6u4H4lFgosWGwtyySIKrEREJ12D3KTxvZhPMLB9YDdxjZneEWNeoUdpzWU7tbBaRsW+woZDn7geBq4AH3H0R0esuj3mRnHSy01M1VkFEksJgQyEtuDjOZ4AnQ6xn1Om+2I5CQUSSwWBD4VvAs8AWd/+Tmc0ANodX1ugSHaugUBCRsW9QF9lx918Dv455vBW4OqyiRpvSSDbvVTcmugwRkdANdkRzmZk9ZmZ7g9sjZlYWdnGjRUkkm32HWnWxHREZ8wbbfXQf8ARQEtx+E7Qlhe7DUvc06AgkERnbBhsKRe5+n7t3BLefAUUh1jWq6ApsIpIsBhsKdWb2OTNLDW6fA+oGWsHMTjKzNTG3g2Z2i5nlm9nzZrY5uJ8YLG9mdqeZVZjZOjNbMNwfbqQcGaugUBCRsW2wofAFooej7gGqgWuAzw+0gru/7+7z3X0+sBBoBh4DbgVedPfZwIvBY4DLgNnBbTlw15B+khBNyeveUlD3kYiMbUM5JPV6dy9y90lEQ+Kfh/A6S4kezroDWAbcH7TfD1wZTC8jOjDO3X0lEAnGRiRcZloqRbrYjogkgcGGwtzYcx25+37gzCG8zrXAQ8H0ZHevDqb3AJOD6VJgV8w6lUHbUcxsuZmtMrNVtbW1QyhheHSxHRFJBoMNhZTuvn+A4BxIgxrjYGYZRM+y+uve8zx6kYIhXajA3e9293J3Ly8qit++7tJIlvYpiMiYN6gvduAHwB/NrPuL/dPAtwe57mXAanevCR7XmFmxu1cH3UN7g/YqYGrMemVB26hQkpfN7zbuxd0xs0SXIyISikFtKbj7A0RPhlcT3K5y958P8jU+y5GuI4iOd7g+mL4eeDym/brgKKTFQENMN1PClUSyaWnv4kBze6JLEREJzWC3FHD3DcCGoTy5meUCHwVuimn+F+BXZnYjsIPoUU0ATxG9ylsF0SOVbhjKa4Ut9roK+bkZCa5GRCQcgw6FY+HuTUBBr7Y6+jjtdrB/4eYw6xmO0phQOL00L8HViIiEY7A7mpOeRjWLSDJQKAxSfm4GmWkp7Nb5j0RkDFMoDFL0YjvZOixVRMY0hcIQ6ApsIjLWKRSGoCRPV2ATkbFNoTAEJZFs9ja20tbRlehSRERCoVAYgtJINu5Qc1A7m0VkbFIoDEGJrqsgImOcQmEINFZBRMY6hcIQxJ7qQkRkLFIoDEFWeioFuRlU6QpsIjJGKRSGqCSiw1JFZOxSKAyRBrCJyFimUBii7i2F6EldRUTGFoXCEJVGsmlq6+Tg4Y5ElyIiMuIUCkOksQoiMpYpFIZIh6WKyFimUBiingFsDQoFERl7FApDVJibSUZqirqPRGRMCjUUzCxiZivMbKOZvWdmZ5vZ7WZWZWZrgtvlMcvfZmYVZva+mV0SZm3HKiXFKI5ksVsD2ERkDEoL+fn/DXjG3a8xswwgB7gEuMPdvx+7oJmdClwLnAaUAC+Y2Rx37wy5xiHTdRVEZKwKbUvBzPKAJcC9AO7e5u71A6yyDPilu7e6+zagAjgrrPqGQ6OaRWSsCrP7aDpQC9xnZm+b2U/MLDeY9yUzW2dmPzWziUFbKbArZv3KoG3UKY1kUXOwhfZOXWxHRMaWMEMhDVgA3OXuZwJNwK3AXcBMYD5QDfxgKE9qZsvNbJWZraqtrR3hkgenJJJNl8OeBu1XEJGxJcxQqAQq3f2N4PEKYIG717h7p7t3AfdwpIuoCpgas35Z0HYUd7/b3cvdvbyoqCjE8vt3SvEEAP64pS4hry8iEpbQQsHd9wC7zOykoGkpsMHMimMW+xSwPph+ArjWzDLNbDowG3gzrPqGY25ZHjMKc1mxujLRpYiIjKiwjz76MvBgcOTRVuAG4E4zmw84sB24CcDd3zWzXwEbgA7g5tF45BGAmXH1wjK+9+z77NrfzNT8nESXJCIyIux4PttneXm5r1q1KiGvvbv+MOd+53d8delsbrl4TkJqEBE5Fmb2lruX9zVPI5qPUUkkm3NmFvDo6iqdRltExgyFwjBcvaCMnfub+dP2A4kuRURkRCgUhuGS06aQk5HKI29ph7OIjA0KhWHIzUzjstOL+e071RxuG5X7xEVEhkShMExXLyzlUGsHz23Yk+hSRESGTaEwTIunF1AayWaFupBEZAxQKAxTSopx1YJSXqvYp9NeiMhxT6EwAq5aUEaXw2Nvf+CsHCIixxWFwgiYXpjLwhMn8sjqSo1ZEJHjmkJhhFy9oIyKvYdYV9mQ6FJERI6ZQmGEfHxuMRlpKTyik+SJyHFMoTBC8rLT+dipk3li7W5aOzRmQUSOTwqFEXT1wjLqm9v5/ca9iS5FROSYKBRG0PmzCikan8mKt3QUkogcnxQKIygtNYVPnVnKH97fS92h1kSXIyIyZAqFEXb1gjI6upzH1+xOdCkiIkOmUBhhJ00Zz+mlE3QUkogclxQKIbh6QRnv7j7Ixj0HE12KiMiQKBRC8Ml5JaSlmK6zICLHHYVCCArGZfKRkyfx2Nu76ejsSnQ5IiKDFmoomFnEzFaY2UYze8/MzjazfDN73sw2B/cTg2XNzO40swozW2dmC8KsLWxXLyhj36FWXtm8L9GliIgMWthbCv8GPOPuJwPzgPeAW4EX3X028GLwGOAyYHZwWw7cFXJtobro5ElMzElnhXY4i8hxJLRQMLM8YAlwL4C7t7l7PbAMuD9Y7H7gymB6GfCAR60EImZWHFZ9YctIS+GT80p4fkMNDc3tiS5HRGRQwtxSmA7UAveZ2dtm9hMzywUmu3t1sMweYHIwXQrsilm/Mmg7bl29sIy2ji6efEdjFkTk+BBmKKQBC4C73P1MoIkjXUUAePTiA0O6AIGZLTezVWa2qra2dsSKDcMZpXnMnjRORyGJyHEjzFCoBCrd/Y3g8QqiIVHT3S0U3HefPa4KmBqzflnQdhR3v9vdy929vKioKLTiR4KZcfXCMlbvrGdr7aFElyMi8qFCCwV33wPsMrOTgqalwAbgCeD6oO164PFg+gnguuAopMVAQ0w303HrU2eWkmLw6GqdJE9ERr+0kJ//y8CDZpYBbAVuIBpEvzKzG4EdwGeCZZ8CLgcqgOZg2ePe5AlZnDe7iMferuLrH51DSooluiQRkX6FGgruvgYo72PW0j6WdeDmMOtJlKsXlPLVX65h5dY6zplVmOhyRET6pRHNcXDJaVMYn5mmMQsiMuopFOIgKz2Vj88t5pn1e2hq7Uh0OSIi/VIoxMnVC8tobuvk6fV7El2KiEi/FApxUn7iRE4syNGYBREZ1RQKcWJmXHVmGX/cWkflgeZElyMi0ieFQhxdtSB61o7HNGZBREYphUIcTc3PYdH0fB59u4roEbgiIqOLQiHOrl5YxrZ9TazeeSDRpYiIfIBCIc4uP6OY7PRUVrylLiQRGX0UCnE2LjONS0+fwpPrdrO/qS3R5YiIHEWhkACfW3wCTa0dXPDd3/PD31fQ3KYBbSIyOigUEmDhifk8c8sSFs8s4HvPvs8F3/sDP1+5g/bOrkSXJiJJTqGQIHMmj+ee68p55G/OZnpBLv/4X+u5+F9f4om1u+nq0pFJIpIYCoUEW3hiPg/ftJj7Pv9nZKen8pWH3uaKf3+VlzbV6rBVEYk7hcIoYGZ85ORJPPWV8/m/fz6fgy3tXP/TN/nsPSt16KqIxJVCYRRJSTGuPLOUF79+IbdfcSqbaw5x1X+8zk0/X0XF3sZElyciScCO5y6K8vJyX7VqVaLLCM2h1g7ufWUb97yylea2Dq5ZWMYtF8+hJJKd6NKO2ZbaQzy5tpo/bt3HyVMmcMGcIhbNyCcnI+yLAIpEuTvrKhsojmQxaXxWostJCDN7y937ugCaQuF4UHeolR/+fgv/uXIHGFx/9on87YWzmJibkejSBmXX/maeXFfNb9buZkP1QczgpMnj2bavidaOLjJSUyifNpElc4pYMruIU4rHY6bLlsrIau/s4jdrd3P3y1vZuKcRM1g0PZ+Pzy3hstOnUDguM9Elxo1CYYyoPNDMHc9v5tG3K8lJT2XJnCIumFPEkjlFo27roeZgC79dV81v1u3m7Z31AMyfGuGKeSV8/IxipuRl0dLeyZ+27+flTbW8vGkf79dEu8gKx2WyZHYh588p5PzZRWPuj9XdWb2znkMxF1zqHYGxmWgxc7vbU8w4vXQC47PSQ6x0bDjU2sEv39zJva9uo7qhhTmTx/H5c6ZTc7CFJ9ftZkttEykGZ88s4ONnlHDp6VPIP07+4TpWCQsFM9sONAKdQIe7l5vZ7cBfA7XBYv/T3Z8Klr8NuDFY/ivu/uxAz59sodDt/T2N3PfaNl7aVEt1QwsAsyeNi/6nPaeIRdPzyUpPjXtddYdaeXr9Hn6zdjdvbt+PO5xaPIFPzCvmirklTM3PGXD9moMtvLypllc27+PVin09I75PK5nA+bOLWDKnkPIT88lIG/quMHentaOLptYOmlo7ae/qYnpBLikp8d0iea1iH999ZiNrKxuG/VzpqcbiGQVcfMpklp4yibKJA7+/yabmYAv3vbadB9/YQWNLB4tn5HPTkplceFJRz5aou/N+TSO/XVfNk+uq2bavidQU45yZBXxibjGXnDaFSM7YC4hEh0K5u++LabsdOOTu3++17KnAQ8BZQAnwAjDH3Tv7e/5kDYVu7s7mvYd4eVMtL22q5Y1t+2nr6CIzLYVFMwq4YE4RF8wpZGbRuNC6YxoOt/Psu9EgeH1LHZ1dzsyiXK6YV8In5pYwa9K4Y3reri5n/e4GXtm8j5c21bJ6xwE6upycjFQWzyjg7BkFpKZY9Eu+rTP4su/gUGsHTW3RL/7Ytua2Tjp6jf+YM3kcf33+DJbNLz2moBmKNbvq+d6zG3mtoo7SSDZfumgWcyYfeW96/xn2/qvsPf9weyevV+zj+fdq2FrbBMDJU8bz0VMnc/EpkzmjNC/ugTdYTa0drK2sZ9L4LGYUjnwwb65p5O6Xt/Jfa6ro7HIuO72Y5UtmMG9qZMD13J0N1Qd7AmLn/mbSUoxzZxXyibnFfOzUKeTljI0ts+MlFG4DcPf/Ezx+Frjd3f/Y3/Mneyj0dritk5Xb6oLumFq2BF8WpZFslswpZMnsIs6ZVUhe9uA/2O7O4fZODrV00Nj9JdvSQXVDC0+v38PLm2pp6+xian42V8wt4Yp5JZw8ZeT3CTS2tLNya9DVtLmWHXVHLlSUkZpCbmYquZlpjMtMIyfjyHRuZhq5wePY+S0dXTy4cgcb9zQyZUIWN5w7jc8uOoEJI9wds6mmke8/+z7PbaihIDeDmz8yi79cfAKZaSO3Jbe19hAvvreX59+rYdX2/XQ5FI3P5OJTJnHxKZM5d1ZhQrYcu7V3drF2Vz2vVuzj9Yo6Vu880BPQ4zPTOL00j3lTI8wri94X52UN+fPj7ry5bT93v7yVFzfuJSs9hc+UT+XG86ZzYkHukGt2d97dfZDfrNvNb9dVU3ngMOmpxvmzi/jE3GIuPnUyE7LSaevo4mBLOw2Hj9wOBrcjjzuOmt9wuJ3GlnbMjOz0VLLSU8hKTyUzPZWstOh0d1tWWq/56SlBWyqnl05gbtnAQdefRIbCNuAA0X98fuzudweh8HngILAK+O/ufsDM/h1Y6e7/Gax7L/C0u6/o7/kVCgOrPNDMy5v28dKmvbxeUUdjawepKcaZUyMsnlGAGTS2HPlvuufWcmS6qbWD/gZYT5mQxcfnFnPFvBLmleXFdefwvkOtpKUYORlpx/xfvrvz8uZ9/PilLby+pY7xmWn8xaITuOHc6UzJG95RKbv2N3PHC5t47O0qxmWk8ddLZvCF86YzLjPco6wONLXxh017eWHDXl7aVMuh1g6y0lM4b1YRHz11EhedPJmi8eHuo+negn118z5eq9jHyq11NLV1YgZnlOZx7qxCzpqez77GVtZW1rOusoH3qg/S3hn9oBWOy2T+1DzmlkWYW5bHvLJIvwdVdHY5z727hx+9vJW1u+rJz83gurNP5Lqzp43YfoHuo5WeDAJid0ML6alGWkoKh9v77cgAICs9hbzsdPKy05mQlX5kOjudLnda2jtpae+K3ndE71u72zo6j5rf2nH0aXD+5sKZfOPSk4/pZ0pkKJS6e5WZTQKeB74MvA/sIxoU/xsodvcvDDYUzGw5sBzghBNOWLhjx47Q6h9L2ju7eHtnfc9/2usqGzCLnrW155Z1ZDq3V3tuZhrjY9ojOemcNHn8qO2iGKp3Khv48ctbeOqdalJTjGXzS1m+ZAZzJo8f0vPsbWzhh7+r4Bdv7iTFjM+fM40vXjAzIUeKtXV08ca2Ol7YUMML7+2lqv4wZjCvLMLSkycxo2gcBeMyKMjNID83g0hOBqnH+PusbjjMaxV1vFYR3R9U29gKwLSCHM6dVch5swo5e2ZBv/3zrR2dvFfdyNpd9T1BsaX2UE+32Qn5OUdtTcwqGsdv36nmJ69sZXtdMycW5PDfzp/BNQvKyM4Ib6uoq8tZU1nP8xtq6Ojs6vmCP+o+q/tx2ohuEXbvF+sOiuz01GPuzhoVRx/11W1kZtOAJ939dHUfxVdbRxfpqaZDP3vZtb+Zn7yylYdX7aKlvYuLTp7E8iUzWDQ9f8D3quFwOz9+aQv3vbadts4u/vzPpvKVi2YPe4tjpLg7G/c0BgFR0+eO7hSDiTnRgIiGRSb5QWAUjssgP3hcOC6DcVlprKts4PUgBLq7KgtyMzhnViHnzSrgnJmFH3pwwUAaW9p5p6qBtbsaWFdZz9pd9ewODqzoNm9qhJuWzOCS06Ycc6Alo4SEgpnlAinu3hhMPw98C1jr7tXBMl8DFrn7tWZ2GvALjuxofhGYrR3NkggHmtr4+cod/Oz17exvauv3y+dwWyf3vb6NH/1hCwdbOvjkvBK+/tE5TCscej92PNU3t7HnYAt1h9qoa2pj/6FW9je1sa+pjf2H2oLpaFt9c3u/z5OdnsqiGfmcN6uQc2cVhr71uLexhXW7Gni/ppHyEydy1oeEtfQtUaEwA3gseJgG/MLdv21mPwfmE+0+2g7cFBMS/wB8AegAbnH3pwd6DYWChK2lvZMVb1Vyzytb2RHTTbFsfgmPv13Fnb+roLaxlYtOnsTffewkTi2ZkOiSR1xHZxcHmtupa2plfxAi9c1tzJk8njNPmBj6kVsy8kZF91EYFAoSL713aKYYdDmcNT2fv7/kJMqn5Se6RJFBGygUdMIZkc3M3VkAAAckSURBVEFITTEuO6OYS0+fwpvb9vPUO9V85ORJXDCnSN0XMqYoFESGwMxYNKOARTMKEl2KSCjUGSgiIj0UCiIi0kOhICIiPRQKIiLSQ6EgIiI9FAoiItJDoSAiIj0UCiIi0uO4Ps2FmdUCx3ru7EKip/AerUZ7fTD6a1R9w6P6hmc013eiuxf1NeO4DoXhMLNV/Z37YzQY7fXB6K9R9Q2P6hue0V5ff9R9JCIiPRQKIiLSI5lD4e5EF/AhRnt9MPprVH3Do/qGZ7TX16ek3acgIiIflMxbCiIi0suYDwUzu9TM3jezCjO7tY/5mWb2cDD/DTObFsfapprZ781sg5m9a2Zf7WOZC82swczWBLd/ild9wetvN7N3gtf+wGXuLOrO4P1bZ2YL4ljbSTHvyxozO2hmt/RaJu7vn5n91Mz2mtn6mLZ8M3vezDYH9xP7Wff6YJnNZnZ9HOv7npltDH6Hj5lZpJ91B/w8hFjf7WZWFfN7vLyfdQf8ew+xvodjattuZmv6WTf092/Y3H3M3oBUYAswA8gA1gKn9lrmb4EfBdPXAg/Hsb5iYEEwPR7Y1Ed9FwJPJvA93A4UDjD/cuBpwIDFwBsJ/F3vIXr8dULfP2AJsABYH9P2XeDWYPpW4Dt9rJcPbA3uJwbTE+NU38eAtGD6O33VN5jPQ4j13Q783SA+AwP+vYdVX6/5PwD+KVHv33BvY31L4Sygwt23unsb8EtgWa9llgH3B9MrgKUWp+srunu1u68OphuB94DSeLz2CFoGPOBRK4GImRUnoI6lwBZ3P9bBjCPG3V8G9vdqjv2c3Q9c2ceqlwDPu/t+dz8APA9cGo/63P05d+8IHq4Eykb6dQern/dvMAbz9z5sA9UXfHd8BnhopF83XsZ6KJQCu2IeV/LBL92eZYI/igYg7tdaDLqtzgTe6GP22Wa21syeNrPT4loYOPCcmb1lZsv7mD+Y9zgerqX/P8REvn/dJrt7dTC9B5jcxzKj5b38AtGtv7582OchTF8Kurd+2k/322h4/84Hatx9cz/zE/n+DcpYD4XjgpmNAx4BbnH3g71mrybaJTIP+H/Af8W5vPPcfQFwGXCzmS2J8+t/KDPLAD4J/LqP2Yl+/z7Ao/0Io/KwPzP7B6ADeLCfRRL1ebgLmAnMB6qJdtGMRp9l4K2EUf/3NNZDoQqYGvO4LGjrcxkzSwPygLq4VBd9zXSigfCguz/ae767H3T3Q8H0U0C6mRXGqz53rwru9wKPEd1EjzWY9zhslwGr3b2m94xEv38xarq71YL7vX0sk9D30sw+D3wC+MsguD5gEJ+HULh7jbt3unsXcE8/r5vo9y8NuAp4uL9lEvX+DcVYD4U/AbPNbHrw3+S1wBO9lnkC6D7K4xrgd/39QYy0oP/xXuA9d//XfpaZ0r2Pw8zOIvo7i0tomVmumY3vnia6M3J9r8WeAK4LjkJaDDTEdJPES7//nSXy/esl9nN2PfB4H8s8C3zMzCYG3SMfC9pCZ2aXAn8PfNLdm/tZZjCfh7Dqi91P9al+Xncwf+9huhjY6O6Vfc1M5Ps3JIne0x32jejRMZuIHpXwD0Hbt4h++AGyiHY7VABvAjPiWNt5RLsR1gFrgtvlwBeBLwbLfAl4l+iRFCuBc+JY34zgddcGNXS/f7H1GfDD4P19ByiP8+83l+iXfF5MW0LfP6IBVQ20E+3XvpHofqoXgc3AC0B+sGw58JOYdb8QfBYrgBviWF8F0f747s9h9xF5JcBTA30e4lTfz4PP1zqiX/TFvesLHn/g7z0e9QXtP+v+3MUsG/f3b7g3jWgWEZEeY737SEREhkChICIiPRQKIiLSQ6EgIiI9FAoiItJDoSCSIMEZXJ9MdB0isRQKIiLSQ6Eg8iHM7HNm9mZwDvwfm1mqmR0yszsseh2MF82sKFh2vpmtjLkuwcSgfZaZvRCcmG+1mc0Mnn6cma0IrmXwYLzO0CvSH4WCyADM7BTgz4Fz3X0+0An8JdGR1Kvc/TTgJeCbwSoPAN9w97lER+B2tz8I/NCjJ+Y7h+iIWIieGfcW4FSiI17PDf2HEhlAWqILEBnllgILgT8F/8RnEz2ZXRdHTnz2n8CjZpYHRNz9paD9fuDXwfluSt39MQB3bwEInu9ND86VE1ytaxrwavg/lkjfFAoiAzPgfne/7ahGs3/stdyxni+mNWa6E/1NSoKp+0hkYC8C15jZJOi51vKJRP92rgmW+QvgVXdvAA6Y2flB+18BL3n0qnqVZnZl8ByZZpYT159CZJD0X4nIANx9g5n9L6JXy0ohembMm4Em4Kxg3l6i+x0gelrsHwVf+luBG4L2vwJ+bGbfCp7j03H8MUQGTWdJFTkGZnbI3cclug6RkabuIxER6aEtBRER6aEtBRER6aFQEBGRHgoFERHpoVAQEZEeCgUREemhUBARkR7/H90TewF4nPsBAAAAAElFTkSuQmCC\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "r59sehdVj0tm",
        "colab_type": "code",
        "outputId": "35d59485-f24d-489f-c876-db1fe78412bf",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 141
        }
      },
      "source": [
        "new_accuracies = []\n",
        "new_examples = []\n",
        "\n",
        "# Run test for each epsilon\n",
        "for eps in epsilons:\n",
        "    nacc, nex = test(model, device, test_loader, eps)\n",
        "    new_accuracies.append(nacc)\n",
        "    new_examples.append(nex)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Epsilon: 0\tTest Accuracy = 6938 / 7000 = 0.9911428571428571\n",
            "Epsilon: 0.05\tTest Accuracy = 6690 / 7000 = 0.9557142857142857\n",
            "Epsilon: 0.1\tTest Accuracy = 6035 / 7000 = 0.8621428571428571\n",
            "Epsilon: 0.15\tTest Accuracy = 4682 / 7000 = 0.6688571428571428\n",
            "Epsilon: 0.2\tTest Accuracy = 2756 / 7000 = 0.39371428571428574\n",
            "Epsilon: 0.25\tTest Accuracy = 1341 / 7000 = 0.19157142857142856\n",
            "Epsilon: 0.3\tTest Accuracy = 661 / 7000 = 0.09442857142857143\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "8KXRlrHdl1HJ",
        "colab_type": "code",
        "outputId": "bbbd8a1a-b1e6-49fa-c5c6-f7f4bec2604c",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 350
        }
      },
      "source": [
        "plt.figure(figsize=(5,5))\n",
        "plt.plot(epsilons, new_accuracies, \"*-\")\n",
        "plt.yticks(np.arange(0, 1.1, step=0.1))\n",
        "plt.xticks(np.arange(0, .35, step=0.05))\n",
        "plt.title(\"Accuracy vs Epsilon\")\n",
        "plt.xlabel(\"Epsilon\")\n",
        "plt.ylabel(\"Accuracy\")\n",
        "plt.show()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 360x360 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    }
  ]
}