CNN_400_Augmented.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "2c33c4cc",
   "metadata": {},
   "outputs": [],
   "source": [
    "import torch\n",
    "import torch.nn as nn\n",
    "import torch.nn.functional as F\n",
    "from torch.utils.data import DataLoader\n",
    "from torchvision import datasets, transforms, models \n",
    "from torchvision.utils import make_grid#\n",
    "import torchvision\n",
    "import os\n",
    "import random\n",
    "from sklearn.model_selection import train_test_split\n",
    "\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline\n",
    "\n",
    "import warnings\n",
    "warnings.filterwarnings(\"ignore\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "2adc55f8",
   "metadata": {},
   "outputs": [],
   "source": [
    "def build_datasets():\n",
    "\n",
    "    transform = transforms.Compose([\n",
    "        transforms.Resize((256,256)),\n",
    "        transforms.ToTensor(),\n",
    "        transforms.Normalize(mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225]),\n",
    "    ])\n",
    "    trainset = torchvision.datasets.ImageFolder(root='Breast_cancer_400A/train', transform=transform)\n",
    "    trainloader = torch.utils.data.DataLoader(trainset, batch_size=32,\n",
    "                                          shuffle=True, num_workers=2)\n",
    "    validset = torchvision.datasets.ImageFolder(root='Breast_cancer_400A/valid', transform=transform)\n",
    "    validloader = torch.utils.data.DataLoader(validset, batch_size=32,\n",
    "                                         shuffle=False, num_workers=2)\n",
    "    testset = torchvision.datasets.ImageFolder(root='Breast_cancer_400A/test', transform=transform)\n",
    "    testloader = torch.utils.data.DataLoader(testset, batch_size=32,\n",
    "                                         shuffle=False, num_workers=2)\n",
    "\n",
    "    return trainloader,validloader, testloader\n",
    "\n",
    "classes = ('Adenosis', 'Ductal_Carcinoma', 'Fibroadenoma', 'Lobular_Carcinoma',\n",
    "               'Mucinous_Carcinoma', 'Papillary_Carcinoma', 'Phyllodes_tumor', 'Tubular_adenoma')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "e3792adb",
   "metadata": {},
   "outputs": [],
   "source": [
    "trainloader,validloader, testloader= build_datasets()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "01e88e25",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "CNN(\n",
       "  (features): Sequential(\n",
       "    (0): Conv2d(3, 96, kernel_size=(11, 11), stride=(4, 4), padding=(2, 2))\n",
       "    (1): ReLU(inplace=True)\n",
       "    (2): MaxPool2d(kernel_size=3, stride=2, padding=0, dilation=1, ceil_mode=False)\n",
       "    (3): Conv2d(96, 256, kernel_size=(5, 5), stride=(1, 1), padding=(2, 2))\n",
       "    (4): ReLU(inplace=True)\n",
       "    (5): MaxPool2d(kernel_size=3, stride=2, padding=0, dilation=1, ceil_mode=False)\n",
       "    (6): Conv2d(256, 384, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))\n",
       "    (7): ReLU(inplace=True)\n",
       "    (8): Conv2d(384, 384, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))\n",
       "    (9): ReLU(inplace=True)\n",
       "    (10): Conv2d(384, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))\n",
       "    (11): ReLU(inplace=True)\n",
       "    (12): MaxPool2d(kernel_size=3, stride=2, padding=0, dilation=1, ceil_mode=False)\n",
       "  )\n",
       "  (maxpool): MaxPool2d(kernel_size=(6, 6), stride=(6, 6), padding=0, dilation=1, ceil_mode=False)\n",
       "  (classifier): Sequential(\n",
       "    (0): Dropout(p=0.5, inplace=False)\n",
       "    (1): Linear(in_features=9216, out_features=4096, bias=True)\n",
       "    (2): ReLU(inplace=True)\n",
       "    (3): Dropout(p=0.5, inplace=False)\n",
       "    (4): Linear(in_features=4096, out_features=1024, bias=True)\n",
       "    (5): ReLU(inplace=True)\n",
       "    (6): Linear(in_features=1024, out_features=8, bias=True)\n",
       "  )\n",
       ")"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import torch.nn as nn\n",
    "import torch.nn.functional as F\n",
    "# define the CNN architecture\n",
    "\n",
    "class CNN(nn.Module):\n",
    "    def __init__(self, num_classes: int = 8, dropout: float = 0.5) -> None:\n",
    "        super().__init__()\n",
    "        self.features = nn.Sequential(\n",
    "            nn.Conv2d(3, 96, kernel_size=11, stride=4, padding=2),\n",
    "            nn.ReLU(inplace=True),\n",
    "            nn.MaxPool2d(kernel_size=3, stride=2),\n",
    "            nn.Conv2d(96, 256, kernel_size=5, padding=2),\n",
    "            nn.ReLU(inplace=True),\n",
    "            nn.MaxPool2d(kernel_size=3, stride=2),\n",
    "            nn.Conv2d(256, 384, kernel_size=3, padding=1),\n",
    "            nn.ReLU(inplace=True),\n",
    "            nn.Conv2d(384, 384, kernel_size=3, padding=1),\n",
    "            nn.ReLU(inplace=True),\n",
    "            nn.Conv2d(384, 256, kernel_size=3, padding=1),\n",
    "            nn.ReLU(inplace=True),\n",
    "            nn.MaxPool2d(kernel_size=3, stride=2),\n",
    "        )\n",
    "        self.maxpool = nn.MaxPool2d((6, 6))\n",
    "        self.classifier = nn.Sequential(\n",
    "            nn.Dropout(p=dropout),\n",
    "            nn.Linear(256 * 6 * 6, 4096),\n",
    "            nn.ReLU(inplace=True),\n",
    "            nn.Dropout(p=dropout),\n",
    "            nn.Linear(4096, 1024),\n",
    "            nn.ReLU(inplace=True),\n",
    "            nn.Linear(1024, num_classes),\n",
    "        )\n",
    "\n",
    "    def forward(self, x: torch.Tensor) -> torch.Tensor:\n",
    "        x = self.features(x)\n",
    "        x = self.maxpool(x)\n",
    "        x = torch.flatten(x, 1)\n",
    "        x = self.classifier(x)\n",
    "        return x\n",
    "\n",
    "net = CNN()\n",
    "net"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "85bd7613",
   "metadata": {},
   "outputs": [],
   "source": [
    "import torch.optim as optim\n",
    "\n",
    "\n",
    "criterion = nn.CrossEntropyLoss()\n",
    "optimizer = optim.SGD(net.parameters(), lr=0.001, momentum=0.9)\n",
    "device = torch.device(\"cuda:0\" if torch.cuda.is_available() else \"cpu\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "b4e2bede",
   "metadata": {},
   "outputs": [],
   "source": [
    "def calculate_accuracy(y_pred, y):\n",
    "    top_pred = y_pred.argmax(1, keepdim = True)\n",
    "    correct = top_pred.eq(y.view_as(top_pred)).sum()\n",
    "    acc = correct.float() / y.shape[0]\n",
    "    return acc\n",
    "\n",
    "def train(model, iterator, optimizer, criterion, device):\n",
    "    epoch_loss = 0\n",
    "    epoch_acc = 0\n",
    "    model.train()\n",
    "    for (x, y) in iterator:\n",
    "        x = x.to(device)\n",
    "        y = y.to(device)\n",
    "\n",
    "        optimizer.zero_grad()\n",
    "\n",
    "        y_pred = model(x)\n",
    "\n",
    "        loss = criterion(y_pred, y)\n",
    "\n",
    "        acc = calculate_accuracy(y_pred, y)\n",
    "\n",
    "        loss.backward()\n",
    "\n",
    "        optimizer.step()\n",
    "\n",
    "        epoch_loss += loss.item()\n",
    "        epoch_acc += acc.item()\n",
    "\n",
    "    return epoch_loss / len(iterator), epoch_acc / len(iterator)\n",
    "\n",
    "def valid(model,iterator):\n",
    "    epoch_loss = 0\n",
    "    epoch_acc = 0\n",
    "\n",
    "    model.eval()\n",
    "\n",
    "    for (x, y) in iterator:\n",
    "        x = x.to(device)\n",
    "        y = y.to(device)\n",
    "        y_pred = model(x)\n",
    "        loss = criterion(y_pred, y)\n",
    "        acc = calculate_accuracy(y_pred, y)\n",
    "\n",
    "        epoch_loss += loss.item()\n",
    "        epoch_acc += acc.item()\n",
    "\n",
    "    return epoch_loss / len(iterator), epoch_acc / len(iterator)\n",
    "\n",
    "\n",
    "def epoch_time(start_time, end_time):\n",
    "    elapsed_time = end_time - start_time\n",
    "    elapsed_mins = int(elapsed_time / 60)\n",
    "    elapsed_secs = int(elapsed_time - (elapsed_mins * 60))\n",
    "    return elapsed_mins, elapsed_secs"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "8d8ba068",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch: 01 | Epoch Time: 3m 23s\n",
      "\tTrain Loss: 1.994 | Train Acc: 18.61%\n",
      "\tValid Loss: 1.922 | Valid Acc: 23.96%\n",
      "Epoch: 02 | Epoch Time: 3m 23s\n",
      "\tTrain Loss: 1.891 | Train Acc: 23.52%\n",
      "\tValid Loss: 1.895 | Valid Acc: 23.75%\n",
      "Epoch: 03 | Epoch Time: 3m 23s\n",
      "\tTrain Loss: 1.828 | Train Acc: 27.09%\n",
      "\tValid Loss: 1.892 | Valid Acc: 28.89%\n",
      "Epoch: 04 | Epoch Time: 3m 23s\n",
      "\tTrain Loss: 1.783 | Train Acc: 30.65%\n",
      "\tValid Loss: 1.847 | Valid Acc: 28.54%\n",
      "Epoch: 05 | Epoch Time: 3m 23s\n",
      "\tTrain Loss: 1.730 | Train Acc: 34.37%\n",
      "\tValid Loss: 1.702 | Valid Acc: 35.83%\n",
      "Epoch: 06 | Epoch Time: 3m 23s\n",
      "\tTrain Loss: 1.667 | Train Acc: 36.00%\n",
      "\tValid Loss: 1.592 | Valid Acc: 36.53%\n",
      "Epoch: 07 | Epoch Time: 3m 23s\n",
      "\tTrain Loss: 1.471 | Train Acc: 44.53%\n",
      "\tValid Loss: 1.748 | Valid Acc: 44.17%\n",
      "Epoch: 08 | Epoch Time: 3m 23s\n",
      "\tTrain Loss: 1.314 | Train Acc: 51.25%\n",
      "\tValid Loss: 1.548 | Valid Acc: 45.90%\n",
      "Epoch: 09 | Epoch Time: 3m 23s\n",
      "\tTrain Loss: 1.084 | Train Acc: 60.54%\n",
      "\tValid Loss: 1.169 | Valid Acc: 67.08%\n",
      "Epoch: 10 | Epoch Time: 3m 22s\n",
      "\tTrain Loss: 0.873 | Train Acc: 68.66%\n",
      "\tValid Loss: 0.904 | Valid Acc: 75.69%\n",
      "Epoch: 11 | Epoch Time: 3m 23s\n",
      "\tTrain Loss: 0.668 | Train Acc: 75.71%\n",
      "\tValid Loss: 0.976 | Valid Acc: 80.21%\n",
      "Epoch: 12 | Epoch Time: 3m 22s\n",
      "\tTrain Loss: 0.496 | Train Acc: 82.92%\n",
      "\tValid Loss: 1.191 | Valid Acc: 79.86%\n",
      "Epoch: 13 | Epoch Time: 3m 23s\n",
      "\tTrain Loss: 0.345 | Train Acc: 88.42%\n",
      "\tValid Loss: 1.010 | Valid Acc: 88.54%\n",
      "Epoch: 14 | Epoch Time: 3m 22s\n",
      "\tTrain Loss: 0.332 | Train Acc: 89.05%\n",
      "\tValid Loss: 1.287 | Valid Acc: 86.67%\n",
      "Epoch: 15 | Epoch Time: 3m 23s\n",
      "\tTrain Loss: 0.232 | Train Acc: 92.74%\n",
      "\tValid Loss: 1.004 | Valid Acc: 81.88%\n",
      "Epoch: 16 | Epoch Time: 3m 22s\n",
      "\tTrain Loss: 0.210 | Train Acc: 93.81%\n",
      "\tValid Loss: 1.356 | Valid Acc: 87.50%\n",
      "Epoch: 17 | Epoch Time: 3m 22s\n",
      "\tTrain Loss: 0.164 | Train Acc: 94.42%\n",
      "\tValid Loss: 0.953 | Valid Acc: 91.04%\n",
      "Epoch: 18 | Epoch Time: 3m 23s\n",
      "\tTrain Loss: 0.143 | Train Acc: 95.95%\n",
      "\tValid Loss: 1.674 | Valid Acc: 87.08%\n",
      "Epoch: 19 | Epoch Time: 3m 22s\n",
      "\tTrain Loss: 0.184 | Train Acc: 94.83%\n",
      "\tValid Loss: 0.977 | Valid Acc: 88.61%\n",
      "Epoch: 20 | Epoch Time: 3m 22s\n",
      "\tTrain Loss: 0.104 | Train Acc: 97.02%\n",
      "\tValid Loss: 1.118 | Valid Acc: 90.62%\n",
      "Epoch: 21 | Epoch Time: 3m 22s\n",
      "\tTrain Loss: 0.158 | Train Acc: 95.62%\n",
      "\tValid Loss: 1.058 | Valid Acc: 88.96%\n",
      "Epoch: 22 | Epoch Time: 3m 23s\n",
      "\tTrain Loss: 0.119 | Train Acc: 96.69%\n",
      "\tValid Loss: 1.060 | Valid Acc: 90.83%\n",
      "Epoch: 23 | Epoch Time: 3m 33s\n",
      "\tTrain Loss: 0.101 | Train Acc: 97.84%\n",
      "\tValid Loss: 1.326 | Valid Acc: 90.21%\n",
      "Epoch: 24 | Epoch Time: 3m 25s\n",
      "\tTrain Loss: 0.041 | Train Acc: 98.80%\n",
      "\tValid Loss: 1.440 | Valid Acc: 88.96%\n",
      "Epoch: 25 | Epoch Time: 3m 42s\n",
      "\tTrain Loss: 0.030 | Train Acc: 99.26%\n",
      "\tValid Loss: 1.687 | Valid Acc: 90.21%\n",
      "Epoch: 26 | Epoch Time: 3m 45s\n",
      "\tTrain Loss: 0.039 | Train Acc: 98.98%\n",
      "\tValid Loss: 1.229 | Valid Acc: 89.79%\n",
      "Epoch: 27 | Epoch Time: 3m 27s\n",
      "\tTrain Loss: 0.039 | Train Acc: 98.78%\n",
      "\tValid Loss: 1.391 | Valid Acc: 90.62%\n",
      "Epoch: 28 | Epoch Time: 3m 22s\n",
      "\tTrain Loss: 0.107 | Train Acc: 97.38%\n",
      "\tValid Loss: 1.117 | Valid Acc: 88.54%\n",
      "Epoch: 29 | Epoch Time: 3m 21s\n",
      "\tTrain Loss: 0.032 | Train Acc: 99.24%\n",
      "\tValid Loss: 1.318 | Valid Acc: 90.42%\n",
      "Epoch: 30 | Epoch Time: 3m 21s\n",
      "\tTrain Loss: 0.056 | Train Acc: 98.47%\n",
      "\tValid Loss: 1.388 | Valid Acc: 91.25%\n",
      "Epoch: 31 | Epoch Time: 3m 22s\n",
      "\tTrain Loss: 0.047 | Train Acc: 98.73%\n",
      "\tValid Loss: 1.628 | Valid Acc: 90.00%\n",
      "Epoch: 32 | Epoch Time: 3m 21s\n",
      "\tTrain Loss: 0.036 | Train Acc: 98.78%\n",
      "\tValid Loss: 1.404 | Valid Acc: 89.38%\n",
      "Epoch: 33 | Epoch Time: 3m 21s\n",
      "\tTrain Loss: 0.019 | Train Acc: 99.47%\n",
      "\tValid Loss: 1.336 | Valid Acc: 91.67%\n",
      "Epoch: 34 | Epoch Time: 3m 21s\n",
      "\tTrain Loss: 0.030 | Train Acc: 99.19%\n",
      "\tValid Loss: 1.486 | Valid Acc: 90.21%\n",
      "Epoch: 35 | Epoch Time: 3m 21s\n",
      "\tTrain Loss: 0.052 | Train Acc: 98.91%\n",
      "\tValid Loss: 1.198 | Valid Acc: 90.42%\n",
      "Epoch: 36 | Epoch Time: 3m 21s\n",
      "\tTrain Loss: 0.054 | Train Acc: 98.65%\n",
      "\tValid Loss: 1.251 | Valid Acc: 90.42%\n",
      "Epoch: 37 | Epoch Time: 3m 21s\n",
      "\tTrain Loss: 0.049 | Train Acc: 98.78%\n",
      "\tValid Loss: 2.061 | Valid Acc: 89.58%\n",
      "Epoch: 38 | Epoch Time: 3m 23s\n",
      "\tTrain Loss: 0.034 | Train Acc: 99.08%\n",
      "\tValid Loss: 1.679 | Valid Acc: 89.58%\n",
      "Epoch: 39 | Epoch Time: 3m 23s\n",
      "\tTrain Loss: 0.030 | Train Acc: 99.13%\n",
      "\tValid Loss: 1.085 | Valid Acc: 91.25%\n",
      "Epoch: 40 | Epoch Time: 3m 21s\n",
      "\tTrain Loss: 0.013 | Train Acc: 99.57%\n",
      "\tValid Loss: 1.521 | Valid Acc: 91.67%\n",
      "Epoch: 41 | Epoch Time: 3m 23s\n",
      "\tTrain Loss: 0.039 | Train Acc: 99.01%\n",
      "\tValid Loss: 1.433 | Valid Acc: 90.83%\n",
      "Epoch: 42 | Epoch Time: 3m 21s\n",
      "\tTrain Loss: 0.028 | Train Acc: 99.36%\n",
      "\tValid Loss: 1.677 | Valid Acc: 90.21%\n",
      "Epoch: 43 | Epoch Time: 3m 22s\n",
      "\tTrain Loss: 0.013 | Train Acc: 99.77%\n",
      "\tValid Loss: 1.496 | Valid Acc: 91.04%\n",
      "Epoch: 44 | Epoch Time: 3m 21s\n",
      "\tTrain Loss: 0.011 | Train Acc: 99.67%\n",
      "\tValid Loss: 1.794 | Valid Acc: 91.25%\n",
      "Epoch: 45 | Epoch Time: 3m 21s\n",
      "\tTrain Loss: 0.021 | Train Acc: 99.36%\n",
      "\tValid Loss: 1.751 | Valid Acc: 91.46%\n",
      "Epoch: 46 | Epoch Time: 3m 22s\n",
      "\tTrain Loss: 0.037 | Train Acc: 99.01%\n",
      "\tValid Loss: 1.631 | Valid Acc: 90.83%\n",
      "Epoch: 47 | Epoch Time: 3m 21s\n",
      "\tTrain Loss: 0.034 | Train Acc: 99.29%\n",
      "\tValid Loss: 1.496 | Valid Acc: 90.62%\n",
      "Epoch: 48 | Epoch Time: 3m 21s\n",
      "\tTrain Loss: 0.020 | Train Acc: 99.59%\n",
      "\tValid Loss: 1.809 | Valid Acc: 90.42%\n",
      "Epoch: 49 | Epoch Time: 3m 22s\n",
      "\tTrain Loss: 0.028 | Train Acc: 99.39%\n",
      "\tValid Loss: 1.927 | Valid Acc: 90.83%\n",
      "Epoch: 50 | Epoch Time: 3m 22s\n",
      "\tTrain Loss: 0.022 | Train Acc: 99.52%\n",
      "\tValid Loss: 1.696 | Valid Acc: 91.25%\n",
      "Epoch: 51 | Epoch Time: 3m 22s\n",
      "\tTrain Loss: 0.013 | Train Acc: 99.64%\n",
      "\tValid Loss: 2.158 | Valid Acc: 90.83%\n",
      "Epoch: 52 | Epoch Time: 3m 21s\n",
      "\tTrain Loss: 0.020 | Train Acc: 99.36%\n",
      "\tValid Loss: 2.120 | Valid Acc: 91.67%\n",
      "Epoch: 53 | Epoch Time: 3m 22s\n",
      "\tTrain Loss: 0.011 | Train Acc: 99.77%\n",
      "\tValid Loss: 1.989 | Valid Acc: 91.04%\n",
      "Epoch: 54 | Epoch Time: 3m 22s\n",
      "\tTrain Loss: 0.039 | Train Acc: 99.08%\n",
      "\tValid Loss: 1.643 | Valid Acc: 90.62%\n",
      "Epoch: 55 | Epoch Time: 3m 21s\n",
      "\tTrain Loss: 0.007 | Train Acc: 99.82%\n",
      "\tValid Loss: 1.977 | Valid Acc: 89.79%\n",
      "Epoch: 56 | Epoch Time: 3m 21s\n",
      "\tTrain Loss: 0.074 | Train Acc: 98.17%\n",
      "\tValid Loss: 1.623 | Valid Acc: 90.00%\n",
      "Epoch: 57 | Epoch Time: 3m 21s\n",
      "\tTrain Loss: 0.153 | Train Acc: 97.45%\n",
      "\tValid Loss: 1.592 | Valid Acc: 90.42%\n",
      "Epoch: 58 | Epoch Time: 3m 23s\n",
      "\tTrain Loss: 0.082 | Train Acc: 98.52%\n",
      "\tValid Loss: 1.408 | Valid Acc: 89.38%\n",
      "Epoch: 59 | Epoch Time: 3m 21s\n",
      "\tTrain Loss: 0.035 | Train Acc: 98.88%\n",
      "\tValid Loss: 1.290 | Valid Acc: 88.96%\n",
      "Epoch: 60 | Epoch Time: 3m 21s\n",
      "\tTrain Loss: 0.041 | Train Acc: 99.21%\n",
      "\tValid Loss: 1.487 | Valid Acc: 91.46%\n",
      "Epoch: 61 | Epoch Time: 3m 22s\n",
      "\tTrain Loss: 0.006 | Train Acc: 99.75%\n",
      "\tValid Loss: 1.582 | Valid Acc: 91.25%\n",
      "Epoch: 62 | Epoch Time: 3m 22s\n",
      "\tTrain Loss: 0.023 | Train Acc: 99.36%\n",
      "\tValid Loss: 2.046 | Valid Acc: 90.42%\n",
      "Epoch: 63 | Epoch Time: 3m 21s\n",
      "\tTrain Loss: 0.004 | Train Acc: 99.87%\n",
      "\tValid Loss: 1.754 | Valid Acc: 91.04%\n",
      "Epoch: 64 | Epoch Time: 3m 22s\n",
      "\tTrain Loss: 0.051 | Train Acc: 98.85%\n",
      "\tValid Loss: 0.943 | Valid Acc: 91.67%\n",
      "Epoch: 65 | Epoch Time: 3m 22s\n",
      "\tTrain Loss: 0.028 | Train Acc: 99.21%\n",
      "\tValid Loss: 1.340 | Valid Acc: 90.83%\n",
      "Epoch: 66 | Epoch Time: 3m 22s\n",
      "\tTrain Loss: 0.016 | Train Acc: 99.64%\n",
      "\tValid Loss: 1.457 | Valid Acc: 91.04%\n",
      "Epoch: 67 | Epoch Time: 3m 22s\n",
      "\tTrain Loss: 0.025 | Train Acc: 99.54%\n",
      "\tValid Loss: 1.309 | Valid Acc: 91.04%\n",
      "Epoch: 68 | Epoch Time: 3m 22s\n",
      "\tTrain Loss: 0.018 | Train Acc: 99.59%\n",
      "\tValid Loss: 1.532 | Valid Acc: 90.21%\n",
      "Epoch: 69 | Epoch Time: 3m 22s\n",
      "\tTrain Loss: 0.014 | Train Acc: 99.77%\n",
      "\tValid Loss: 1.740 | Valid Acc: 91.67%\n",
      "Epoch: 70 | Epoch Time: 3m 22s\n",
      "\tTrain Loss: 0.006 | Train Acc: 99.85%\n",
      "\tValid Loss: 1.976 | Valid Acc: 91.46%\n",
      "Epoch: 71 | Epoch Time: 3m 23s\n",
      "\tTrain Loss: 0.007 | Train Acc: 99.80%\n",
      "\tValid Loss: 1.528 | Valid Acc: 90.62%\n",
      "Epoch: 72 | Epoch Time: 3m 23s\n",
      "\tTrain Loss: 0.008 | Train Acc: 99.77%\n",
      "\tValid Loss: 1.670 | Valid Acc: 89.58%\n",
      "Epoch: 73 | Epoch Time: 3m 22s\n",
      "\tTrain Loss: 0.032 | Train Acc: 99.16%\n",
      "\tValid Loss: 1.600 | Valid Acc: 91.25%\n",
      "Epoch: 74 | Epoch Time: 3m 23s\n",
      "\tTrain Loss: 0.018 | Train Acc: 99.54%\n",
      "\tValid Loss: 1.545 | Valid Acc: 90.62%\n",
      "Epoch: 75 | Epoch Time: 3m 22s\n",
      "\tTrain Loss: 0.023 | Train Acc: 99.49%\n",
      "\tValid Loss: 1.440 | Valid Acc: 90.83%\n",
      "Epoch: 76 | Epoch Time: 3m 23s\n",
      "\tTrain Loss: 0.005 | Train Acc: 99.87%\n",
      "\tValid Loss: 1.526 | Valid Acc: 91.25%\n",
      "Epoch: 77 | Epoch Time: 3m 24s\n",
      "\tTrain Loss: 0.007 | Train Acc: 99.75%\n",
      "\tValid Loss: 1.390 | Valid Acc: 91.25%\n",
      "Epoch: 78 | Epoch Time: 3m 22s\n",
      "\tTrain Loss: 0.002 | Train Acc: 99.95%\n",
      "\tValid Loss: 1.624 | Valid Acc: 91.04%\n",
      "Epoch: 79 | Epoch Time: 3m 22s\n",
      "\tTrain Loss: 0.003 | Train Acc: 99.90%\n",
      "\tValid Loss: 1.747 | Valid Acc: 91.04%\n",
      "Epoch: 80 | Epoch Time: 3m 22s\n",
      "\tTrain Loss: 0.002 | Train Acc: 99.95%\n",
      "\tValid Loss: 1.912 | Valid Acc: 90.83%\n",
      "Epoch: 81 | Epoch Time: 3m 21s\n",
      "\tTrain Loss: 0.021 | Train Acc: 99.49%\n",
      "\tValid Loss: 1.636 | Valid Acc: 90.62%\n",
      "Epoch: 82 | Epoch Time: 3m 24s\n",
      "\tTrain Loss: 0.005 | Train Acc: 99.82%\n",
      "\tValid Loss: 1.861 | Valid Acc: 90.83%\n",
      "Epoch: 83 | Epoch Time: 3m 24s\n",
      "\tTrain Loss: 0.004 | Train Acc: 99.87%\n",
      "\tValid Loss: 1.797 | Valid Acc: 91.25%\n",
      "Epoch: 84 | Epoch Time: 3m 21s\n",
      "\tTrain Loss: 0.001 | Train Acc: 100.00%\n",
      "\tValid Loss: 1.971 | Valid Acc: 91.67%\n",
      "Epoch: 85 | Epoch Time: 3m 22s\n",
      "\tTrain Loss: 0.007 | Train Acc: 99.77%\n",
      "\tValid Loss: 2.227 | Valid Acc: 90.62%\n",
      "Epoch: 86 | Epoch Time: 3m 22s\n",
      "\tTrain Loss: 0.003 | Train Acc: 99.92%\n",
      "\tValid Loss: 1.898 | Valid Acc: 91.25%\n",
      "Epoch: 87 | Epoch Time: 3m 25s\n",
      "\tTrain Loss: 0.002 | Train Acc: 99.92%\n",
      "\tValid Loss: 2.063 | Valid Acc: 91.46%\n",
      "Epoch: 88 | Epoch Time: 3m 22s\n",
      "\tTrain Loss: 0.004 | Train Acc: 99.92%\n",
      "\tValid Loss: 1.988 | Valid Acc: 91.46%\n",
      "Epoch: 89 | Epoch Time: 3m 23s\n",
      "\tTrain Loss: 0.002 | Train Acc: 99.95%\n",
      "\tValid Loss: 2.111 | Valid Acc: 91.67%\n",
      "Epoch: 90 | Epoch Time: 3m 22s\n",
      "\tTrain Loss: 0.001 | Train Acc: 99.95%\n",
      "\tValid Loss: 2.120 | Valid Acc: 91.67%\n",
      "Epoch: 91 | Epoch Time: 3m 22s\n",
      "\tTrain Loss: 0.030 | Train Acc: 99.44%\n",
      "\tValid Loss: 1.964 | Valid Acc: 90.83%\n",
      "Epoch: 92 | Epoch Time: 3m 22s\n",
      "\tTrain Loss: 0.007 | Train Acc: 99.75%\n",
      "\tValid Loss: 1.505 | Valid Acc: 90.83%\n",
      "Epoch: 93 | Epoch Time: 3m 22s\n",
      "\tTrain Loss: 0.004 | Train Acc: 99.85%\n",
      "\tValid Loss: 2.114 | Valid Acc: 91.04%\n",
      "Epoch: 94 | Epoch Time: 3m 23s\n",
      "\tTrain Loss: 0.002 | Train Acc: 99.97%\n",
      "\tValid Loss: 2.212 | Valid Acc: 90.83%\n",
      "Epoch: 95 | Epoch Time: 3m 23s\n",
      "\tTrain Loss: 0.016 | Train Acc: 99.64%\n",
      "\tValid Loss: 2.116 | Valid Acc: 90.42%\n",
      "Epoch: 96 | Epoch Time: 3m 23s\n",
      "\tTrain Loss: 0.019 | Train Acc: 99.59%\n",
      "\tValid Loss: 1.661 | Valid Acc: 90.83%\n",
      "Epoch: 97 | Epoch Time: 3m 23s\n",
      "\tTrain Loss: 0.008 | Train Acc: 99.77%\n",
      "\tValid Loss: 1.808 | Valid Acc: 91.04%\n",
      "Epoch: 98 | Epoch Time: 3m 23s\n",
      "\tTrain Loss: 0.012 | Train Acc: 99.77%\n",
      "\tValid Loss: 1.624 | Valid Acc: 90.42%\n",
      "Epoch: 99 | Epoch Time: 3m 22s\n",
      "\tTrain Loss: 0.004 | Train Acc: 99.92%\n",
      "\tValid Loss: 1.747 | Valid Acc: 91.04%\n",
      "Epoch: 100 | Epoch Time: 3m 22s\n",
      "\tTrain Loss: 0.009 | Train Acc: 99.87%\n",
      "\tValid Loss: 2.086 | Valid Acc: 90.83%\n",
      "Finished Training\n",
      "Accuracy of the network test images: 90 %\n"
     ]
    }
   ],
   "source": [
    "import seaborn as sn\n",
    "import torch.nn as nn\n",
    "import torch.nn.functional as F\n",
    "import time\n",
    "\n",
    "train_l = []\n",
    "train_a = []\n",
    "valid_l = []\n",
    "valid_a = []\n",
    "\n",
    "if __name__ == '__main__':\n",
    "\n",
    "    for epoch in range(100):\n",
    "        start_time = time.monotonic()\n",
    "        train_loss, train_acc = train(net, trainloader, optimizer, criterion, device)\n",
    "        valid_loss, valid_acc = valid(net, validloader)\n",
    "        \n",
    "        train_l.append(train_loss)\n",
    "        train_a.append(train_acc)\n",
    "        valid_l.append(valid_loss)\n",
    "        valid_a.append(valid_acc)\n",
    "        \n",
    "        end_time = time.monotonic()\n",
    "        epoch_mins, epoch_secs = epoch_time(start_time, end_time)\n",
    "\n",
    "        print(f'Epoch: {epoch + 1:02} | Epoch Time: {epoch_mins}m {epoch_secs}s')\n",
    "        print(f'\\tTrain Loss: {train_loss:.3f} | Train Acc: {train_acc * 100:.2f}%')\n",
    "        print(f'\\tValid Loss: {valid_loss:.3f} | Valid Acc: {valid_acc * 100:.2f}%')\n",
    "    print('Finished Training')\n",
    "\n",
    "    correct = 0\n",
    "    total = 0\n",
    "    with torch.no_grad():\n",
    "        for data in testloader:\n",
    "            images, labels = data\n",
    "            outputs = net(images)\n",
    "            _, predicted = torch.max(outputs.data, 1)\n",
    "            total += labels.size(0)\n",
    "            correct += (predicted == labels).sum().item()\n",
    "\n",
    "    print('Accuracy of the network test images: %d %%' % (\n",
    "            100 * correct / total))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "50c2f9c6",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[0.18609979633401222, 0.23523421588594703, 0.2708757637474542, 0.3065173116089613, 0.3436863543788187, 0.359979633401222, 0.44526476578411406, 0.5124745417515275, 0.6053971486761711, 0.6866089613034623, 0.7571283095723014, 0.8291751527494908, 0.8841649694501018, 0.8905295315682281, 0.9274439918533605, 0.9381364562118126, 0.9442464358452138, 0.9595213849287169, 0.9483197556008147, 0.970213849287169, 0.9562118126272913, 0.9669042769857433, 0.9783604887983707, 0.9880346232179226, 0.9926171079429735, 0.9898167006109979, 0.9877800407331976, 0.9737780040733197, 0.9923625254582484, 0.984725050916497, 0.9872708757637475, 0.9877800407331976, 0.994653767820774, 0.9918533604887984, 0.9890529531568228, 0.9865071283095723, 0.9877800407331976, 0.9908350305498982, 0.9913441955193483, 0.9956720977596741, 0.990071283095723, 0.9936354378818737, 0.9977087576374746, 0.9966904276985743, 0.9936354378818737, 0.990071283095723, 0.9928716904276986, 0.9959266802443992, 0.9938900203665988, 0.995162932790224, 0.9964358452138493, 0.9936354378818737, 0.9977087576374746, 0.9908350305498982, 0.9982179226069247, 0.9816700610997964, 0.9745417515274949, 0.9852342158859471, 0.9887983706720977, 0.9921079429735234, 0.9974541751527495, 0.9936354378818737, 0.9987270875763747, 0.9885437881873728, 0.9921079429735234, 0.9964358452138493, 0.995417515274949, 0.9959266802443992, 0.9977087576374746, 0.9984725050916496, 0.9979633401221996, 0.9977087576374746, 0.9915987780040734, 0.995417515274949, 0.994908350305499, 0.9987270875763747, 0.9974541751527495, 0.9994908350305499, 0.9989816700610998, 0.9994908350305499, 0.994908350305499, 0.9982179226069247, 0.9987270875763747, 1.0, 0.9977087576374746, 0.9992362525458248, 0.9992362525458248, 0.9992362525458248, 0.9994908350305499, 0.9994908350305499, 0.9943991853360489, 0.9974541751527495, 0.9984725050916496, 0.9997454175152749, 0.9964358452138493, 0.9959266802443992, 0.9977087576374746, 0.9977087576374746, 0.9992362525458248, 0.9987270875763747]\n"
     ]
    }
   ],
   "source": [
    "print(train_a)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "3d5512c5",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[0.23958333333333334, 0.2375, 0.28888888905445737, 0.28541666666666665, 0.35833333333333334, 0.3652777781089147, 0.44166666666666665, 0.4590277781089147, 0.6708333333333333, 0.7569444447755813, 0.8020833333333334, 0.7986111114422481, 0.8854166666666666, 0.8666666666666667, 0.81875, 0.875, 0.9104166666666667, 0.8708333333333333, 0.8861111114422481, 0.90625, 0.8895833333333333, 0.9083333333333333, 0.9020833333333333, 0.8895833333333333, 0.9020833333333333, 0.8979166666666667, 0.90625, 0.8854166666666666, 0.9041666666666667, 0.9125, 0.9, 0.89375, 0.9166666666666666, 0.9020833333333333, 0.9041666666666667, 0.9041666666666667, 0.8958333333333334, 0.8958333333333334, 0.9125, 0.9166666666666666, 0.9083333333333333, 0.9020833333333333, 0.9104166666666667, 0.9125, 0.9145833333333333, 0.9083333333333333, 0.90625, 0.9041666666666667, 0.9083333333333333, 0.9125, 0.9083333333333333, 0.9166666666666666, 0.9104166666666667, 0.90625, 0.8979166666666667, 0.9, 0.9041666666666667, 0.89375, 0.8895833333333333, 0.9145833333333333, 0.9125, 0.9041666666666667, 0.9104166666666667, 0.9166666666666666, 0.9083333333333333, 0.9104166666666667, 0.9104166666666667, 0.9020833333333333, 0.9166666666666666, 0.9145833333333333, 0.90625, 0.8958333333333334, 0.9125, 0.90625, 0.9083333333333333, 0.9125, 0.9125, 0.9104166666666667, 0.9104166666666667, 0.9083333333333333, 0.90625, 0.9083333333333333, 0.9125, 0.9166666666666666, 0.90625, 0.9125, 0.9145833333333333, 0.9145833333333333, 0.9166666666666666, 0.9166666666666666, 0.9083333333333333, 0.9083333333333333, 0.9104166666666667, 0.9083333333333333, 0.9041666666666667, 0.9083333333333333, 0.9104166666666667, 0.9041666666666667, 0.9104166666666667, 0.9083333333333333]\n"
     ]
    }
   ],
   "source": [
    "print(valid_a)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "343632a0",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy for class Adenosis is: 84.7 %\n",
      "Accuracy for class Ductal_Carcinoma is: 93.3 %\n",
      "Accuracy for class Fibroadenoma is: 84.5 %\n",
      "Accuracy for class Lobular_Carcinoma is: 90.9 %\n",
      "Accuracy for class Mucinous_Carcinoma is: 96.6 %\n",
      "Accuracy for class Papillary_Carcinoma is: 89.8 %\n",
      "Accuracy for class Phyllodes_tumor is: 90.8 %\n",
      "Accuracy for class Tubular_adenoma is: 95.1 %\n"
     ]
    }
   ],
   "source": [
    "classes = ('Adenosis', 'Ductal_Carcinoma', 'Fibroadenoma', 'Lobular_Carcinoma',\n",
    "               'Mucinous_Carcinoma', 'Papillary_Carcinoma', 'Phyllodes_tumor', 'Tubular_adenoma')\n",
    "\n",
    "# prepare to count predictions for each class\n",
    "correct_pred = {classname: 0 for classname in classes}\n",
    "total_pred = {classname: 0 for classname in classes}\n",
    "\n",
    "# again no gradients needed\n",
    "with torch.no_grad():\n",
    "    for data in testloader:\n",
    "        images, labels = data\n",
    "        outputs = net(images)\n",
    "        _, predictions = torch.max(outputs, 1)\n",
    "        # collect the correct predictions for each class\n",
    "        for label, prediction in zip(labels, predictions):\n",
    "            if label == prediction:\n",
    "                correct_pred[classes[label]] += 1\n",
    "            total_pred[classes[label]] += 1\n",
    "\n",
    "\n",
    "# print accuracy for each class\n",
    "for classname, correct_count in correct_pred.items():\n",
    "    accuracy = 100 * float(correct_count) / total_pred[classname]\n",
    "    print(\"Accuracy for class {:5s} is: {:.1f} %\".format(classname,\n",
    "                                                   accuracy))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "00b5bbe6",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Confusion matrix:\n",
      "[[50  0  3  0  0  2  4  0]\n",
      " [ 0 56  0  0  4  0  0  0]\n",
      " [ 1  7 49  0  0  0  1  0]\n",
      " [ 0  3  1 50  0  1  0  0]\n",
      " [ 0  1  0  1 56  0  0  0]\n",
      " [ 0  6  0  0  0 53  0  0]\n",
      " [ 1  2  0  0  3  0 59  0]\n",
      " [ 0  3  0  0  0  0  0 58]]\n",
      "F1 score: 0.907368\n",
      "Accuracy score: 0.907368\n"
     ]
    }
   ],
   "source": [
    "import sklearn\n",
    "\n",
    "def test_label_predictions(model, device, test_loader):\n",
    "    model.eval()\n",
    "    actuals = []\n",
    "    predictions = []\n",
    "    with torch.no_grad():\n",
    "        for data, target in test_loader:\n",
    "            data, target = data.to(device), target.to(device)\n",
    "            output = model(data)\n",
    "            prediction = output.argmax(dim=1, keepdim=True)\n",
    "            actuals.extend(target.view_as(prediction))\n",
    "            predictions.extend(prediction)\n",
    "    return [i.item() for i in actuals], [i.item() for i in predictions]\n",
    "\n",
    "actuals, predictions = test_label_predictions(net, device, testloader)\n",
    "print('Confusion matrix:')\n",
    "print(sklearn.metrics.confusion_matrix(actuals, predictions))\n",
    "print('F1 score: %f' % sklearn.metrics.f1_score(actuals, predictions, average='micro'))\n",
    "print('Accuracy score: %f' % sklearn.metrics.accuracy_score(actuals, predictions))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "5a0e17e6",
   "metadata": {},
   "outputs": [],
   "source": [
    "a = [[50/59,  0,  3/59,  0,  0,  2/59,  4/59,  0],\n",
    "     [ 0 ,56/59 , 0 , 0 , 4/59 , 0,  0,  0],\n",
    "     [ 1/58 , 7/58 ,49/58 , 0  ,0 , 0  ,1/58 , 0],\n",
    "     [ 0 , 3/55 , 1/55, 50/55,  0,  1/55,  0 , 0],\n",
    "     [ 0 , 1/58 , 0 , 1/58 ,53/58  ,3/58  ,0 , 0],\n",
    "     [ 0 , 6/59,  0,  0,  0, 53/59,  0 , 0],\n",
    "     [ 1/65  ,2/65 , 0 , 0  ,3/65 , 0, 59/65 , 0],\n",
    "     [ 0 , 3/62,  0 , 0,  0 , 2/62,  1/62, 55/62]]\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "91b7c46d",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "03bea4ff",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<AxesSubplot:>"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x504 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "classes = ('Adenosis', 'Ductal_Carcinoma', 'Fibroadenoma', 'Lobular_Carcinoma',\n",
    "               'Mucinous_Carcinoma', 'Papillary_Carcinoma', 'Phyllodes_tumor', 'Tubular_adenoma')\n",
    "\n",
    "import seaborn as sn\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "array = a\n",
    "df_cm = pd.DataFrame(array, index = [i for i in classes],\n",
    "                  columns = [i for i in classes])\n",
    "plt.figure(figsize = (10,7))\n",
    "sn.heatmap(df_cm, annot=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "dbd094f1",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "def test_class_probabilities(model, device, test_loader, which_class):\n",
    "    model.eval()\n",
    "    actuals = []\n",
    "    probabilities = []\n",
    "    with torch.no_grad():\n",
    "        for data, target in test_loader:\n",
    "            data, target = data.to(device), target.to(device)\n",
    "            output = model(data)\n",
    "            prediction = output.argmax(dim=1, keepdim=True)\n",
    "            actuals.extend(target.view_as(prediction) == which_class)\n",
    "            probabilities.extend(np.exp(output[:, which_class]))\n",
    "    return [i.item() for i in actuals], [i.item() for i in probabilities]\n",
    "\n",
    "actuals, class_probabilities = test_class_probabilities(net, device, testloader, 0)\n",
    "actuals1, class_probabilities1 = test_class_probabilities(net, device, testloader, 1)\n",
    "actuals2, class_probabilities2 = test_class_probabilities(net, device, testloader, 2)\n",
    "actuals3, class_probabilities3 = test_class_probabilities(net, device, testloader, 3)\n",
    "actuals4, class_probabilities4 = test_class_probabilities(net, device, testloader, 4)\n",
    "actuals5, class_probabilities5 = test_class_probabilities(net, device, testloader, 5)\n",
    "actuals6, class_probabilities6 = test_class_probabilities(net, device, testloader, 6)\n",
    "actuals7, class_probabilities7 = test_class_probabilities(net, device, testloader, 7)\n",
    "\n",
    "\n",
    "fpr, tpr, _ = sklearn.metrics.roc_curve(actuals, np.nan_to_num(class_probabilities))\n",
    "fpr1, tpr1, _ = sklearn.metrics.roc_curve(actuals1, np.nan_to_num(class_probabilities1))\n",
    "fpr2, tpr2, _ = sklearn.metrics.roc_curve(actuals2, np.nan_to_num(class_probabilities2))\n",
    "fpr3, tpr3, _ = sklearn.metrics.roc_curve(actuals3, np.nan_to_num(class_probabilities3))\n",
    "fpr4, tpr4, _ = sklearn.metrics.roc_curve(actuals4, np.nan_to_num(class_probabilities4))\n",
    "fpr5, tpr5, _ = sklearn.metrics.roc_curve(actuals5, np.nan_to_num(class_probabilities5))\n",
    "fpr6, tpr6, _ = sklearn.metrics.roc_curve(actuals6, np.nan_to_num(class_probabilities6))\n",
    "fpr7, tpr7, _ = sklearn.metrics.roc_curve(actuals7, np.nan_to_num(class_probabilities7))\n",
    "\n",
    "roc_auc = sklearn.metrics.auc(fpr, tpr)\n",
    "roc_auc1 = sklearn.metrics.auc(fpr1, tpr1)\n",
    "roc_auc2 = sklearn.metrics.auc(fpr2, tpr2)\n",
    "roc_auc3 = sklearn.metrics.auc(fpr3, tpr3)\n",
    "roc_auc4 = sklearn.metrics.auc(fpr4, tpr4)\n",
    "roc_auc5 = sklearn.metrics.auc(fpr5, tpr5)\n",
    "roc_auc6 = sklearn.metrics.auc(fpr6, tpr6)\n",
    "roc_auc7 = sklearn.metrics.auc(fpr7, tpr7)\n",
    "\n",
    "\n",
    "plt.figure(figsize=(10,5))\n",
    "lw = 2\n",
    "plt.plot(fpr, tpr, label='ROC curve Adenosis (area = %0.2f)' % roc_auc)\n",
    "plt.plot(fpr1, tpr1,lw=lw,label='ROC curve Ductal_Carcinoma (area = %0.2f)' % roc_auc1)\n",
    "plt.plot(fpr2, tpr2, lw=lw,label='ROC curve Fibroadenoma (area = %0.2f)' % roc_auc2)\n",
    "plt.plot(fpr3, tpr3, lw=lw,label='ROC curve Lobular_Carcinoma (area = %0.2f)' % roc_auc3)\n",
    "plt.plot(fpr4, tpr4, lw=lw,label='ROC curve Mucinous_Carcinoma (area = %0.2f)' % roc_auc4)\n",
    "plt.plot(fpr5, tpr5, lw=lw,label='ROC curve Papillary_Carcinoma (area = %0.2f)' % roc_auc5)\n",
    "plt.plot(fpr6, tpr6, lw=lw,label='ROC curve Phyllodes_tumor (area = %0.2f)' % roc_auc6)\n",
    "plt.plot(fpr7, tpr7, lw=lw,label='ROC curve Tubular_adenoma (area = %0.2f)' % roc_auc7)\n",
    "\n",
    "plt.plot([0, 1], [0, 1], color='navy', lw=lw, linestyle='--')\n",
    "plt.xlim([0.0, 1.0])\n",
    "plt.ylim([0.0, 1.05])\n",
    "plt.xlabel('False Positive Rate')\n",
    "plt.ylabel('True Positive Rate')\n",
    "plt.legend(loc=\"lower right\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "7bb99371",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "aaa714e2",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "6ca8ce30",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.8"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}