Classification_Pytorch_TL.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "import torch\n",
    "import numpy as np\n",
    "from torch import nn\n",
    "from torch import optim\n",
    "import os\n",
    "\n",
    "import torch.nn.functional as F\n",
    "from torchvision import datasets ,transforms, models\n",
    "%matplotlib inline\n",
    "%config InlineBackend.figure_format = 'retina'\n",
    "import matplotlib.pyplot as plt\n",
    "from torch.autograd import Variable\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['with_mask', 'without_mask']\n",
      "torch.Size([64, 3, 224, 224])\n",
      "torch.Size([64, 3, 224, 224])\n",
      "torch.Size([64, 3, 224, 224])\n",
      "torch.Size([64, 3, 224, 224])\n",
      "torch.Size([64, 3, 224, 224])\n",
      "torch.Size([64, 3, 224, 224])\n",
      "torch.Size([64, 3, 224, 224])\n",
      "torch.Size([64, 3, 224, 224])\n",
      "torch.Size([64, 3, 224, 224])\n",
      "torch.Size([64, 3, 224, 224])\n",
      "torch.Size([64, 3, 224, 224])\n",
      "torch.Size([64, 3, 224, 224])\n",
      "torch.Size([64, 3, 224, 224])\n",
      "torch.Size([64, 3, 224, 224])\n",
      "torch.Size([64, 3, 224, 224])\n",
      "torch.Size([64, 3, 224, 224])\n",
      "torch.Size([64, 3, 224, 224])\n",
      "torch.Size([37, 3, 224, 224])\n"
     ]
    }
   ],
   "source": [
    "data_dir= 'data/train2'\n",
    "\n",
    "def load_split_train_test(datadir,valid_size=.2):\n",
    "    # create a tranformation pipeline , all the images go \n",
    "    #under all the tranformations\n",
    "    \n",
    "    train_transforms=transforms.Compose([transforms.Resize((224,224)),transforms.ToTensor(),])\n",
    "    test_transforms=transforms.Compose([transforms.Resize((224,224)),transforms.ToTensor(),])\n",
    "    \n",
    "    # callable funtions for loadeing data from image forlders,\n",
    "    #apllying transformatioon\n",
    "    train_data=  datasets.ImageFolder(datadir, transform=train_transforms)\n",
    "    test_data=  datasets.ImageFolder(datadir,transform=test_transforms)\n",
    "    \n",
    "    # spliting train and test indices\n",
    "    num_train= len(train_data)\n",
    "    indices=list(range(num_train))\n",
    "    split=int(np.floor(valid_size*num_train))\n",
    "    \n",
    "    np.random.shuffle(indices)\n",
    "    from torch.utils.data.sampler import SubsetRandomSampler\n",
    "    train_idx, test_idx=indices[split:],indices[:split]\n",
    "    \n",
    "    train_sampler= SubsetRandomSampler(train_idx)\n",
    "    test_sampler= SubsetRandomSampler(test_idx)\n",
    "    \n",
    "    trainloader=torch.utils.data.DataLoader(train_data,sampler=train_sampler, \n",
    "                                            batch_size=64)\n",
    "    testloader=torch.utils.data.DataLoader(test_data,sampler=train_sampler, \n",
    "                                            batch_size=64)\n",
    "    \n",
    "    return trainloader,testloader\n",
    "\n",
    "trainloader,testloader=load_split_train_test(data_dir,.2)\n",
    "\n",
    "print(trainloader.dataset.classes)\n",
    "\n",
    "for batch_images, targets in trainloader:\n",
    "    print(batch_images.shape)  # should be 32*3*224*224\n",
    "\n",
    "\n",
    "    \n",
    "    \n",
    "    \n",
    "    \n",
    "    \n",
    "    \n",
    "    \n",
    "    \n",
    "    \n",
    "    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "cuda\n",
      "ResNet(\n",
      "  (conv1): Conv2d(3, 64, kernel_size=(7, 7), stride=(2, 2), padding=(3, 3), bias=False)\n",
      "  (bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "  (relu): ReLU(inplace=True)\n",
      "  (maxpool): MaxPool2d(kernel_size=3, stride=2, padding=1, dilation=1, ceil_mode=False)\n",
      "  (layer1): Sequential(\n",
      "    (0): Bottleneck(\n",
      "      (conv1): Conv2d(64, 64, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "      (bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "      (conv2): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "      (bn2): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "      (conv3): Conv2d(64, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "      (bn3): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "      (relu): ReLU(inplace=True)\n",
      "      (downsample): Sequential(\n",
      "        (0): Conv2d(64, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "        (1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "      )\n",
      "    )\n",
      "    (1): Bottleneck(\n",
      "      (conv1): Conv2d(256, 64, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "      (bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "      (conv2): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "      (bn2): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "      (conv3): Conv2d(64, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "      (bn3): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "      (relu): ReLU(inplace=True)\n",
      "    )\n",
      "    (2): Bottleneck(\n",
      "      (conv1): Conv2d(256, 64, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "      (bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "      (conv2): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "      (bn2): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "      (conv3): Conv2d(64, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "      (bn3): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "      (relu): ReLU(inplace=True)\n",
      "    )\n",
      "  )\n",
      "  (layer2): Sequential(\n",
      "    (0): Bottleneck(\n",
      "      (conv1): Conv2d(256, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "      (bn1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "      (conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)\n",
      "      (bn2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "      (conv3): Conv2d(128, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "      (bn3): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "      (relu): ReLU(inplace=True)\n",
      "      (downsample): Sequential(\n",
      "        (0): Conv2d(256, 512, kernel_size=(1, 1), stride=(2, 2), bias=False)\n",
      "        (1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "      )\n",
      "    )\n",
      "    (1): Bottleneck(\n",
      "      (conv1): Conv2d(512, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "      (bn1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "      (conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "      (bn2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "      (conv3): Conv2d(128, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "      (bn3): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "      (relu): ReLU(inplace=True)\n",
      "    )\n",
      "    (2): Bottleneck(\n",
      "      (conv1): Conv2d(512, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "      (bn1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "      (conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "      (bn2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "      (conv3): Conv2d(128, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "      (bn3): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "      (relu): ReLU(inplace=True)\n",
      "    )\n",
      "    (3): Bottleneck(\n",
      "      (conv1): Conv2d(512, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "      (bn1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "      (conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "      (bn2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "      (conv3): Conv2d(128, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "      (bn3): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "      (relu): ReLU(inplace=True)\n",
      "    )\n",
      "  )\n",
      "  (layer3): Sequential(\n",
      "    (0): Bottleneck(\n",
      "      (conv1): Conv2d(512, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "      (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "      (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)\n",
      "      (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "      (conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "      (bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "      (relu): ReLU(inplace=True)\n",
      "      (downsample): Sequential(\n",
      "        (0): Conv2d(512, 1024, kernel_size=(1, 1), stride=(2, 2), bias=False)\n",
      "        (1): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "      )\n",
      "    )\n",
      "    (1): Bottleneck(\n",
      "      (conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "      (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "      (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "      (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "      (conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "      (bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "      (relu): ReLU(inplace=True)\n",
      "    )\n",
      "    (2): Bottleneck(\n",
      "      (conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "      (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "      (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "      (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "      (conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "      (bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "      (relu): ReLU(inplace=True)\n",
      "    )\n",
      "    (3): Bottleneck(\n",
      "      (conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "      (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "      (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "      (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "      (conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "      (bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "      (relu): ReLU(inplace=True)\n",
      "    )\n",
      "    (4): Bottleneck(\n",
      "      (conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "      (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "      (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "      (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "      (conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "      (bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "      (relu): ReLU(inplace=True)\n",
      "    )\n",
      "    (5): Bottleneck(\n",
      "      (conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "      (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "      (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "      (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "      (conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "      (bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "      (relu): ReLU(inplace=True)\n",
      "    )\n",
      "  )\n",
      "  (layer4): Sequential(\n",
      "    (0): Bottleneck(\n",
      "      (conv1): Conv2d(1024, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "      (bn1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "      (conv2): Conv2d(512, 512, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)\n",
      "      (bn2): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "      (conv3): Conv2d(512, 2048, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "      (bn3): BatchNorm2d(2048, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "      (relu): ReLU(inplace=True)\n",
      "      (downsample): Sequential(\n",
      "        (0): Conv2d(1024, 2048, kernel_size=(1, 1), stride=(2, 2), bias=False)\n",
      "        (1): BatchNorm2d(2048, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "      )\n",
      "    )\n",
      "    (1): Bottleneck(\n",
      "      (conv1): Conv2d(2048, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "      (bn1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "      (conv2): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "      (bn2): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "      (conv3): Conv2d(512, 2048, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "      (bn3): BatchNorm2d(2048, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "      (relu): ReLU(inplace=True)\n",
      "    )\n",
      "    (2): Bottleneck(\n",
      "      (conv1): Conv2d(2048, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "      (bn1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "      (conv2): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "      (bn2): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "      (conv3): Conv2d(512, 2048, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "      (bn3): BatchNorm2d(2048, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "      (relu): ReLU(inplace=True)\n",
      "    )\n",
      "  )\n",
      "  (avgpool): AdaptiveAvgPool2d(output_size=(1, 1))\n",
      "  (fc): Linear(in_features=2048, out_features=1000, bias=True)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      ")\n"
     ]
    }
   ],
   "source": [
    "# check if gpu is present or not \n",
    "device=torch.device(\"cuda\" if torch.cuda.is_available() else \"cpu\")\n",
    "print(device)\n",
    "model= models.resnet50(pretrained=True)\n",
    "print(model)\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "os.environ['CUDA_LAUNCH_BLOCKING']='0'\n",
    "torch.cuda.get_device_name(torch.cuda.current_device())\n",
    "torch.cuda.set_device(0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "ResNet(\n",
       "  (conv1): Conv2d(3, 64, kernel_size=(7, 7), stride=(2, 2), padding=(3, 3), bias=False)\n",
       "  (bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (relu): ReLU(inplace=True)\n",
       "  (maxpool): MaxPool2d(kernel_size=3, stride=2, padding=1, dilation=1, ceil_mode=False)\n",
       "  (layer1): Sequential(\n",
       "    (0): Bottleneck(\n",
       "      (conv1): Conv2d(64, 64, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "      (bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      (conv2): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "      (bn2): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      (conv3): Conv2d(64, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "      (bn3): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      (relu): ReLU(inplace=True)\n",
       "      (downsample): Sequential(\n",
       "        (0): Conv2d(64, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      )\n",
       "    )\n",
       "    (1): Bottleneck(\n",
       "      (conv1): Conv2d(256, 64, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "      (bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      (conv2): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "      (bn2): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      (conv3): Conv2d(64, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "      (bn3): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      (relu): ReLU(inplace=True)\n",
       "    )\n",
       "    (2): Bottleneck(\n",
       "      (conv1): Conv2d(256, 64, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "      (bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      (conv2): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "      (bn2): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      (conv3): Conv2d(64, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "      (bn3): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      (relu): ReLU(inplace=True)\n",
       "    )\n",
       "  )\n",
       "  (layer2): Sequential(\n",
       "    (0): Bottleneck(\n",
       "      (conv1): Conv2d(256, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "      (bn1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      (conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)\n",
       "      (bn2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      (conv3): Conv2d(128, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "      (bn3): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      (relu): ReLU(inplace=True)\n",
       "      (downsample): Sequential(\n",
       "        (0): Conv2d(256, 512, kernel_size=(1, 1), stride=(2, 2), bias=False)\n",
       "        (1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      )\n",
       "    )\n",
       "    (1): Bottleneck(\n",
       "      (conv1): Conv2d(512, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "      (bn1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      (conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "      (bn2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      (conv3): Conv2d(128, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "      (bn3): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      (relu): ReLU(inplace=True)\n",
       "    )\n",
       "    (2): Bottleneck(\n",
       "      (conv1): Conv2d(512, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "      (bn1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      (conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "      (bn2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      (conv3): Conv2d(128, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "      (bn3): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      (relu): ReLU(inplace=True)\n",
       "    )\n",
       "    (3): Bottleneck(\n",
       "      (conv1): Conv2d(512, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "      (bn1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      (conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "      (bn2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      (conv3): Conv2d(128, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "      (bn3): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      (relu): ReLU(inplace=True)\n",
       "    )\n",
       "  )\n",
       "  (layer3): Sequential(\n",
       "    (0): Bottleneck(\n",
       "      (conv1): Conv2d(512, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "      (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)\n",
       "      (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      (conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "      (bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      (relu): ReLU(inplace=True)\n",
       "      (downsample): Sequential(\n",
       "        (0): Conv2d(512, 1024, kernel_size=(1, 1), stride=(2, 2), bias=False)\n",
       "        (1): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      )\n",
       "    )\n",
       "    (1): Bottleneck(\n",
       "      (conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "      (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "      (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      (conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "      (bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      (relu): ReLU(inplace=True)\n",
       "    )\n",
       "    (2): Bottleneck(\n",
       "      (conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "      (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "      (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      (conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "      (bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      (relu): ReLU(inplace=True)\n",
       "    )\n",
       "    (3): Bottleneck(\n",
       "      (conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "      (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "      (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      (conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "      (bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      (relu): ReLU(inplace=True)\n",
       "    )\n",
       "    (4): Bottleneck(\n",
       "      (conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "      (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "      (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      (conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "      (bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      (relu): ReLU(inplace=True)\n",
       "    )\n",
       "    (5): Bottleneck(\n",
       "      (conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "      (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "      (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      (conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "      (bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      (relu): ReLU(inplace=True)\n",
       "    )\n",
       "  )\n",
       "  (layer4): Sequential(\n",
       "    (0): Bottleneck(\n",
       "      (conv1): Conv2d(1024, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "      (bn1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      (conv2): Conv2d(512, 512, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)\n",
       "      (bn2): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      (conv3): Conv2d(512, 2048, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "      (bn3): BatchNorm2d(2048, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      (relu): ReLU(inplace=True)\n",
       "      (downsample): Sequential(\n",
       "        (0): Conv2d(1024, 2048, kernel_size=(1, 1), stride=(2, 2), bias=False)\n",
       "        (1): BatchNorm2d(2048, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      )\n",
       "    )\n",
       "    (1): Bottleneck(\n",
       "      (conv1): Conv2d(2048, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "      (bn1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      (conv2): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "      (bn2): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      (conv3): Conv2d(512, 2048, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "      (bn3): BatchNorm2d(2048, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      (relu): ReLU(inplace=True)\n",
       "    )\n",
       "    (2): Bottleneck(\n",
       "      (conv1): Conv2d(2048, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "      (bn1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      (conv2): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "      (bn2): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      (conv3): Conv2d(512, 2048, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "      (bn3): BatchNorm2d(2048, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      (relu): ReLU(inplace=True)\n",
       "    )\n",
       "  )\n",
       "  (avgpool): AdaptiveAvgPool2d(output_size=(1, 1))\n",
       "  (fc): Sequential(\n",
       "    (0): Linear(in_features=2048, out_features=512, bias=True)\n",
       "    (1): ReLU()\n",
       "    (2): Dropout(p=0.2, inplace=False)\n",
       "    (3): Linear(in_features=512, out_features=10, bias=True)\n",
       "    (4): LogSoftmax()\n",
       "  )\n",
       ")"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "for param in model.parameters():\n",
    "    param.requires_grad=False\n",
    "# New layer for classification\n",
    "model.fc=nn.Sequential(nn.Linear(2048,512),\n",
    "                       nn.ReLU(),\n",
    "                       nn.Dropout(0.2),\n",
    "                       nn.Linear(512,10),\n",
    "                       nn.LogSoftmax(dim=1))\n",
    "\n",
    "criterion=nn.NLLLoss()\n",
    "optimizer=optim.Adam(model.fc.parameters(),lr=0.003)\n",
    "\n",
    "model.to(device)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/10.. Train loss: 0.755.. Test loss: 0.434.. Test accuracy: 0.821\n",
      "Saving Model after 0 epochs\n",
      "Epoch 2/10.. Train loss: 0.250.. Test loss: 0.054.. Test accuracy: 0.985\n",
      "Epoch 2/10.. Train loss: 0.111.. Test loss: 0.065.. Test accuracy: 0.978\n",
      "Epoch 3/10.. Train loss: 0.081.. Test loss: 0.056.. Test accuracy: 0.982\n",
      "Epoch 3/10.. Train loss: 0.073.. Test loss: 0.048.. Test accuracy: 0.983\n",
      "Saving Model after 2 epochs\n",
      "Epoch 4/10.. Train loss: 0.036.. Test loss: 0.054.. Test accuracy: 0.982\n",
      "Epoch 4/10.. Train loss: 0.043.. Test loss: 0.029.. Test accuracy: 0.993\n",
      "Epoch 5/10.. Train loss: 0.026.. Test loss: 0.029.. Test accuracy: 0.989\n",
      "Epoch 5/10.. Train loss: 0.036.. Test loss: 0.043.. Test accuracy: 0.985\n",
      "Saving Model after 4 epochs\n",
      "Epoch 6/10.. Train loss: 0.022.. Test loss: 0.025.. Test accuracy: 0.992\n",
      "Epoch 7/10.. Train loss: 0.024.. Test loss: 0.025.. Test accuracy: 0.989\n",
      "Epoch 7/10.. Train loss: 0.028.. Test loss: 0.031.. Test accuracy: 0.990\n",
      "Saving Model after 6 epochs\n",
      "Epoch 8/10.. Train loss: 0.019.. Test loss: 0.021.. Test accuracy: 0.996\n",
      "Epoch 8/10.. Train loss: 0.022.. Test loss: 0.017.. Test accuracy: 0.997\n",
      "Epoch 9/10.. Train loss: 0.020.. Test loss: 0.018.. Test accuracy: 0.992\n",
      "Epoch 9/10.. Train loss: 0.017.. Test loss: 0.027.. Test accuracy: 0.992\n",
      "Saving Model after 8 epochs\n",
      "Epoch 10/10.. Train loss: 0.027.. Test loss: 0.036.. Test accuracy: 0.986\n",
      "Epoch 10/10.. Train loss: 0.021.. Test loss: 0.023.. Test accuracy: 0.995\n"
     ]
    }
   ],
   "source": [
    "\n",
    "epochs=10\n",
    "steps=0\n",
    "\n",
    "running_loss=0\n",
    "\n",
    "print_every=10\n",
    "\n",
    "train_losses,test_losses=[],[]\n",
    "\n",
    "for epoch in range(epochs):\n",
    "    for inputs,labels in trainloader:\n",
    "        steps+=1\n",
    "        inputs,labels=inputs.to(device),labels.to(device)\n",
    "        optimizer.zero_grad()\n",
    "        logps=model.forward(inputs)\n",
    "        loss=criterion(logps,labels)\n",
    "        loss.backward()\n",
    "        optimizer.step()\n",
    "        running_loss+=loss.item()\n",
    "        \n",
    "        if steps%print_every==0:\n",
    "            test_loss=0\n",
    "            accuracy=0\n",
    "            for inputs,labels in testloader:\n",
    "                inputs,labels=inputs.to(device),labels.to(device)\n",
    "                logps=model.forward(inputs)\n",
    "                batch_loss=criterion(logps,labels)\n",
    "                test_loss += batch_loss.item()\n",
    "                ps=torch.exp(logps)\n",
    "                \n",
    "                top_p,top_class=ps.topk(1,dim=1)\n",
    "                equals= top_class==labels.view(*top_class.shape)\n",
    "                accuracy+=torch.mean(equals.type(torch.FloatTensor)).item()\n",
    "            train_losses.append(running_loss/len(trainloader))\n",
    "            test_losses.append(test_loss/len(testloader))\n",
    "            \n",
    "            print(f\"Epoch {epoch+1}/{epochs}.. \"\n",
    "                  f\"Train loss: {running_loss/print_every:.3f}.. \"\n",
    "                  f\"Test loss: {test_loss/len(testloader):.3f}.. \"\n",
    "                  f\"Test accuracy: {accuracy/len(testloader):.3f}\")\n",
    "            running_loss=0\n",
    "            model.train()\n",
    "            \n",
    "    if epoch%2==0:\n",
    "        print(\"Saving Model after \" +str(epoch) +\" epochs\")\n",
    "        torch.save(model,'classifymodel_6.pth')\n",
    "\n",
    "torch.save(model,'classifymodel_6.pth')\n",
    "            "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "image/png": {
       "height": 248,
       "width": 377
      },
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(train_losses,label='Traning Loss')\n",
    "plt.plot(test_losses,label='Validation Loss')\n",
    "\n",
    "plt.legend(frameon=False)\n",
    "plt.show()\n"
   ]
  }
 ],
 "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": "3.7.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}