NaturalImages.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import os\n",
    "from random import shuffle\n",
    "import torch\n",
    "import torch.nn as nn\n",
    "from torch.autograd import Variable\n",
    "import torch.utils.data as data\n",
    "import torch.optim as optim\n",
    "from torch.optim import lr_scheduler\n",
    "import torchvision.models as models\n",
    "from torchvision import transforms\n",
    "from PIL import Image\n",
    "from tqdm import tqdm"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Using GPU\n"
     ]
    }
   ],
   "source": [
    "device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')\n",
    "\n",
    "if torch.cuda.is_available():\n",
    "    use_gpu = True\n",
    "    print(\"Using GPU\")\n",
    "else:\n",
    "    use_gpu = False\n",
    "FloatTensor = torch.cuda.FloatTensor if use_gpu else torch.FloatTensor\n",
    "LongTensor = torch.cuda.LongTensor if use_gpu else torch.LongTensor\n",
    "ByteTensor = torch.cuda.ByteTensor if use_gpu else torch.ByteTensor\n",
    "Tensor = FloatTensor"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "class ImageDataset(data.Dataset):\n",
    "    \n",
    "    def __init__(self, image_paths, class_dict, transform=None):\n",
    "        self.image_paths = image_paths\n",
    "        self.transform = transform\n",
    "        self.class_dict = class_dict\n",
    "    \n",
    "    def __len__(self):\n",
    "        return len(self.image_paths)\n",
    "    \n",
    "    def __getitem__(self, index):\n",
    "        img_path = self.image_paths[index]\n",
    "        image = Image.open(img_path)\n",
    "        if self.transform:\n",
    "            image = self.transform(image)\n",
    "        for key, value in class_dict.items():\n",
    "            if key in img_path:\n",
    "                label = value\n",
    "                break\n",
    "        label = torch.tensor(label)\n",
    "        return [image, label]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_resnet_model(num_classes):\n",
    "    resnet = models.resnet18(pretrained='imagenet')\n",
    "    num_hidden = resnet.fc.in_features\n",
    "    resnet.fc = nn.Linear(num_hidden, num_classes)\n",
    "    for name, params in resnet.named_parameters():\n",
    "        if 'layer3' not in name:\n",
    "            params.requires_grad = False\n",
    "        else:\n",
    "            break\n",
    "    return resnet"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "def train_model(model, dataloaders, dataset_sizes, criterion, optimizer, scheduler, num_epochs=25):\n",
    "    best_model_wts = model.state_dict()\n",
    "    best_acc = 0.0\n",
    "\n",
    "    for epoch in range(num_epochs):\n",
    "        print('Epoch {}/{}'.format(epoch, num_epochs - 1))\n",
    "        print('-' * 10)\n",
    "\n",
    "        for phase in ['train', 'val']:\n",
    "            \n",
    "            if phase == 'train':\n",
    "                scheduler.step()\n",
    "                model.train(True)\n",
    "            else:\n",
    "                model.train(False)\n",
    "\n",
    "            running_loss = 0.0\n",
    "            running_corrects = 0.0\n",
    "\n",
    "            for data in tqdm(dataloaders[phase]):\n",
    "                inputs, labels = data\n",
    "\n",
    "                inputs = Variable(inputs.type(Tensor))\n",
    "                labels = Variable(labels.type(LongTensor))\n",
    "\n",
    "                optimizer.zero_grad()\n",
    "\n",
    "                outputs = model(inputs)\n",
    "                if type(outputs) == tuple:\n",
    "                    outputs, _ = outputs\n",
    "                _, preds = torch.max(outputs.data, 1)\n",
    "                loss = criterion(outputs, labels)\n",
    "\n",
    "                if phase == 'train':\n",
    "                    loss.backward()\n",
    "                    optimizer.step()\n",
    "                \n",
    "                batch_loss = loss.data[0]\n",
    "                batch_acc = torch.sum(preds == labels.data)\n",
    "                \n",
    "                running_loss += batch_loss\n",
    "                running_corrects += batch_acc\n",
    "\n",
    "            epoch_loss = running_loss / dataset_sizes[phase]\n",
    "            epoch_acc = float(running_corrects) / dataset_sizes[phase]\n",
    "\n",
    "            print('{} Loss: {:.4f} Acc: {:.4f}'.format(\n",
    "                phase, epoch_loss, epoch_acc))\n",
    "\n",
    "            if phase == 'val' and epoch_acc > best_acc:\n",
    "                best_acc = epoch_acc\n",
    "                best_model_wts = model.state_dict()\n",
    "\n",
    "        print()\n",
    "\n",
    "    print('Best validation accuarcy: {:4f}'.format(best_acc))\n",
    "\n",
    "    model.load_state_dict(best_model_wts)\n",
    "    return model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "data_dir = './natural_images/'\n",
    "mean = [0.485, 0.456, 0.406]\n",
    "std = [0.229, 0.224, 0.225]\n",
    "scale = 360\n",
    "input_shape = 224\n",
    "batch_size = 128\n",
    "epochs = 20\n",
    "num_classes = 8"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "img_files = []\n",
    "classes = os.listdir(data_dir)\n",
    "class_dict = {label: i for i, label in enumerate(classes)}\n",
    "\n",
    "for directory in os.listdir(data_dir):\n",
    "    for filename in os.listdir(os.path.join(data_dir, directory)):\n",
    "        img_files.append(os.path.join(data_dir, directory, filename))\n",
    "\n",
    "shuffle(img_files)\n",
    "\n",
    "train_length = int(0.6*len(img_files))\n",
    "val_length = int(0.8*len(img_files))\n",
    "img_files = {'train': img_files[:train_length],\n",
    "             'val': img_files[train_length:val_length],\n",
    "             'test': img_files[val_length:]}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "data_transforms = {\n",
    "    'train': transforms.Compose([\n",
    "            transforms.Resize(scale),\n",
    "            transforms.RandomResizedCrop(input_shape),\n",
    "            transforms.RandomHorizontalFlip(),\n",
    "            transforms.RandomVerticalFlip(),\n",
    "            transforms.ToTensor(),\n",
    "            transforms.Normalize(mean, std)]),\n",
    "    \n",
    "    'val': transforms.Compose([\n",
    "        transforms.Resize(scale),\n",
    "        transforms.CenterCrop(input_shape),\n",
    "        transforms.ToTensor(),\n",
    "        transforms.Normalize(mean, std)])\n",
    "}\n",
    "\n",
    "\n",
    "image_datasets = {x: ImageDataset(img_files[x], class_dict, data_transforms[x]) for x in ['train', 'val']}\n",
    "\n",
    "dataloaders = {x: torch.utils.data.DataLoader(image_datasets[x], batch_size=batch_size,\n",
    "                                         shuffle=True, num_workers=4) for x in ['train', 'val']}\n",
    "\n",
    "dataset_sizes = {x: float(len(image_datasets[x])) for x in ['train', 'val']}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r",
      "  0%|          | 0/33 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 0/19\n",
      "----------\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/usr/local/lib/python3.5/dist-packages/ipykernel_launcher.py:38: UserWarning: invalid index of a 0-dim tensor. This will be an error in PyTorch 0.5. Use tensor.item() to convert a 0-dim tensor to a Python number\n",
      "100%|██████████| 33/33 [00:16<00:00,  1.95it/s]\n",
      "  0%|          | 0/11 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train Loss: 0.0098 Acc: 0.6564\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 11/11 [00:03<00:00,  2.83it/s]\n",
      "  0%|          | 0/33 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "val Loss: 0.0032 Acc: 0.9225\n",
      "\n",
      "Epoch 1/19\n",
      "----------\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 33/33 [00:17<00:00,  1.94it/s]\n",
      "  0%|          | 0/11 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train Loss: 0.0024 Acc: 0.9394\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 11/11 [00:04<00:00,  2.69it/s]\n",
      "  0%|          | 0/33 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "val Loss: 0.0014 Acc: 0.9659\n",
      "\n",
      "Epoch 2/19\n",
      "----------\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 33/33 [00:17<00:00,  1.93it/s]\n",
      "  0%|          | 0/11 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train Loss: 0.0015 Acc: 0.9524\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 11/11 [00:03<00:00,  2.80it/s]\n",
      "  0%|          | 0/33 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "val Loss: 0.0010 Acc: 0.9725\n",
      "\n",
      "Epoch 3/19\n",
      "----------\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 33/33 [00:17<00:00,  1.89it/s]\n",
      "  0%|          | 0/11 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train Loss: 0.0012 Acc: 0.9633\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 11/11 [00:03<00:00,  2.78it/s]\n",
      "  0%|          | 0/33 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "val Loss: 0.0009 Acc: 0.9761\n",
      "\n",
      "Epoch 4/19\n",
      "----------\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 33/33 [00:17<00:00,  1.86it/s]\n",
      "  0%|          | 0/11 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train Loss: 0.0010 Acc: 0.9667\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 11/11 [00:04<00:00,  2.73it/s]\n",
      "  0%|          | 0/33 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "val Loss: 0.0008 Acc: 0.9790\n",
      "\n",
      "Epoch 5/19\n",
      "----------\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 33/33 [00:17<00:00,  1.92it/s]\n",
      "  0%|          | 0/11 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train Loss: 0.0010 Acc: 0.9642\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 11/11 [00:04<00:00,  2.74it/s]\n",
      "  0%|          | 0/33 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "val Loss: 0.0007 Acc: 0.9812\n",
      "\n",
      "Epoch 6/19\n",
      "----------\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 33/33 [00:17<00:00,  1.88it/s]\n",
      "  0%|          | 0/11 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train Loss: 0.0008 Acc: 0.9681\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 11/11 [00:04<00:00,  2.25it/s]\n",
      "  0%|          | 0/33 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "val Loss: 0.0007 Acc: 0.9790\n",
      "\n",
      "Epoch 7/19\n",
      "----------\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 33/33 [00:17<00:00,  1.91it/s]\n",
      "  0%|          | 0/11 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train Loss: 0.0009 Acc: 0.9705\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 11/11 [00:03<00:00,  2.81it/s]\n",
      "  0%|          | 0/33 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "val Loss: 0.0006 Acc: 0.9812\n",
      "\n",
      "Epoch 8/19\n",
      "----------\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 33/33 [00:17<00:00,  1.87it/s]\n",
      "  0%|          | 0/11 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train Loss: 0.0008 Acc: 0.9710\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 11/11 [00:04<00:00,  2.35it/s]\n",
      "  0%|          | 0/33 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "val Loss: 0.0006 Acc: 0.9812\n",
      "\n",
      "Epoch 9/19\n",
      "----------\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 33/33 [00:17<00:00,  1.88it/s]\n",
      "  0%|          | 0/11 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train Loss: 0.0008 Acc: 0.9686\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 11/11 [00:03<00:00,  2.75it/s]\n",
      "  0%|          | 0/33 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "val Loss: 0.0006 Acc: 0.9804\n",
      "\n",
      "Epoch 10/19\n",
      "----------\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 33/33 [00:17<00:00,  1.85it/s]\n",
      "  0%|          | 0/11 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train Loss: 0.0008 Acc: 0.9751\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 11/11 [00:04<00:00,  2.60it/s]\n",
      "  0%|          | 0/33 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "val Loss: 0.0006 Acc: 0.9790\n",
      "\n",
      "Epoch 11/19\n",
      "----------\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 33/33 [00:17<00:00,  1.87it/s]\n",
      "  0%|          | 0/11 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train Loss: 0.0008 Acc: 0.9737\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 11/11 [00:04<00:00,  2.75it/s]\n",
      "  0%|          | 0/33 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "val Loss: 0.0006 Acc: 0.9790\n",
      "\n",
      "Epoch 12/19\n",
      "----------\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 33/33 [00:17<00:00,  1.85it/s]\n",
      "  0%|          | 0/11 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train Loss: 0.0008 Acc: 0.9717\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 11/11 [00:04<00:00,  2.62it/s]\n",
      "  0%|          | 0/33 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "val Loss: 0.0006 Acc: 0.9790\n",
      "\n",
      "Epoch 13/19\n",
      "----------\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 33/33 [00:17<00:00,  1.88it/s]\n",
      "  0%|          | 0/11 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train Loss: 0.0008 Acc: 0.9667\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 11/11 [00:04<00:00,  2.39it/s]\n",
      "  0%|          | 0/33 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "val Loss: 0.0006 Acc: 0.9797\n",
      "\n",
      "Epoch 14/19\n",
      "----------\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 33/33 [00:17<00:00,  1.87it/s]\n",
      "  0%|          | 0/11 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train Loss: 0.0008 Acc: 0.9732\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 11/11 [00:04<00:00,  2.74it/s]\n",
      "  0%|          | 0/33 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "val Loss: 0.0006 Acc: 0.9812\n",
      "\n",
      "Epoch 15/19\n",
      "----------\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 33/33 [00:17<00:00,  1.87it/s]\n",
      "  0%|          | 0/11 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train Loss: 0.0008 Acc: 0.9729\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 11/11 [00:03<00:00,  2.80it/s]\n",
      "  0%|          | 0/33 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "val Loss: 0.0006 Acc: 0.9797\n",
      "\n",
      "Epoch 16/19\n",
      "----------\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 33/33 [00:17<00:00,  1.88it/s]\n",
      "  0%|          | 0/11 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train Loss: 0.0008 Acc: 0.9715\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 11/11 [00:04<00:00,  2.72it/s]\n",
      "  0%|          | 0/33 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "val Loss: 0.0006 Acc: 0.9790\n",
      "\n",
      "Epoch 17/19\n",
      "----------\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 33/33 [00:17<00:00,  1.86it/s]\n",
      "  0%|          | 0/11 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train Loss: 0.0008 Acc: 0.9725\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 11/11 [00:04<00:00,  2.71it/s]\n",
      "  0%|          | 0/33 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "val Loss: 0.0006 Acc: 0.9797\n",
      "\n",
      "Epoch 18/19\n",
      "----------\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 33/33 [00:17<00:00,  1.89it/s]\n",
      "  0%|          | 0/11 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train Loss: 0.0007 Acc: 0.9720\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 11/11 [00:04<00:00,  2.50it/s]\n",
      "  0%|          | 0/33 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "val Loss: 0.0006 Acc: 0.9790\n",
      "\n",
      "Epoch 19/19\n",
      "----------\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 33/33 [00:17<00:00,  1.85it/s]\n",
      "  0%|          | 0/11 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train Loss: 0.0007 Acc: 0.9722\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 11/11 [00:04<00:00,  2.66it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "val Loss: 0.0006 Acc: 0.9797\n",
      "\n",
      "Best validation accuarcy: 0.981159\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    }
   ],
   "source": [
    "model = get_resnet_model(num_classes).to(device)\n",
    "\n",
    "criterion = nn.CrossEntropyLoss()\n",
    "optimizer = optim.SGD(list(filter(lambda p: p.requires_grad, model.parameters())), lr=0.001, momentum=0.9)\n",
    "exp_lr_scheduler = lr_scheduler.StepLR(optimizer, step_size=7, gamma=0.1)\n",
    "\n",
    "model = train_model(model, dataloaders, dataset_sizes, criterion, optimizer, exp_lr_scheduler, num_epochs=epochs)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "test_transform = transforms.Compose([\n",
    "        transforms.Resize(scale),\n",
    "        transforms.CenterCrop(input_shape),\n",
    "        transforms.ToTensor(),\n",
    "        transforms.Normalize(mean, std)])\n",
    "\n",
    "test_set = ImageDataset(img_files['test'], class_dict, test_transform)\n",
    "\n",
    "test_loader = torch.utils.data.DataLoader(test_set, batch_size=batch_size,\n",
    "                                         shuffle=True, num_workers=4)\n",
    "\n",
    "test_size = float(len(test_set))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 11/11 [00:04<00:00,  2.59it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Test Accuracy: 0.9855072463768116\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    }
   ],
   "source": [
    "model.train(False)\n",
    "\n",
    "running_loss = 0.0\n",
    "running_corrects = 0.0\n",
    "\n",
    "for data in tqdm(test_loader):\n",
    "    inputs, labels = data\n",
    "\n",
    "    inputs = Variable(inputs.type(Tensor))\n",
    "    labels = Variable(labels.type(LongTensor))\n",
    "\n",
    "    outputs = model(inputs)\n",
    "    if type(outputs) == tuple:\n",
    "        outputs, _ = outputs\n",
    "    _, preds = torch.max(outputs.data, 1)\n",
    "    \n",
    "    running_corrects += torch.sum(preds == labels.data)\n",
    "\n",
    "accuracy = float(running_corrects) / test_size\n",
    "print(\"Test Accuracy: {}\".format(accuracy))"
   ]
  }
 ],
 "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": "3.5.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}