{"cells":[{"metadata":{"_uuid":"8f2839f25d086af736a60e9eeb907d3b93b6e0e5","_cell_guid":"b1076dfc-b9ad-4769-8c92-a6c4dae69d19","trusted":true},"cell_type":"code","source":"# This Python 3 environment comes with many helpful analytics libraries installed\n# It is defined by the kaggle/python Docker image: https://github.com/kaggle/docker-python\n# For example, here's several helpful packages to load\n\nimport numpy as np # linear algebra\nimport pandas as pd # data processing, CSV file I/O (e.g. pd.read_csv)\n\n# Input data files are available in the read-only \"../input/\" directory\n# For example, running this (by clicking run or pressing Shift+Enter) will list all files under the input directory\n\nimport os\nfor dirname, _, filenames in os.walk('/kaggle/input'):\n    for filename in filenames:\n        print(os.path.join(dirname, filename))\n\n# You can write up to 5GB to the current directory (/kaggle/working/) that gets preserved as output when you create a version using \"Save & Run All\" \n# You can also write temporary files to /kaggle/temp/, but they won't be saved outside of the current session","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"# import zipfile\n# train_zip_file = '../input/dogs-vs-cats-redux-kernels-edition/train.zip'\n# test_zip_file = '../input/dogs-vs-cats-redux-kernels-edition/test.zip'\n# with zipfile.ZipFile(train_zip_file,\"r\") as z:\n#     z.extractall(\".\")\n# with zipfile.ZipFile(test_zip_file,\"r\") as z:\n#     z.extractall(\".\")","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"# ! pip install loguru","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"! ls /kaggle/working","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"import torch\nimport torch.nn as nn\nimport torch.optim as optim\nfrom torch.utils.data import DataLoader, Dataset\nfrom torchvision import transforms\nimport os\nfrom PIL import Image\nimport matplotlib.pyplot as plt\nimport numpy as np\nimport torchvision\nimport pandas as pd\nfrom loguru import logger\n\nlogger.remove()\nlogger.add(\"./logs/file_{time}.log\", )\n\ntrain_dir = '/kaggle/working/train'\ntest_dir = '/kaggle/working/test'\n\nclass DogCat(Dataset):\n    def __init__(self, root, train=True, test=False, transform=None):\n        self.root = root\n        self.train = train\n        self.test = test\n        self.transform = transform\n        images = [ os.path.join(root, image) for image in os.listdir(root)]\n        if self.test:\n            self.images = images\n        elif train:\n            images_len = len(images)\n            self.images = images[:int(0.8*images_len)]\n        else:\n            images_len = len(images)\n            self.images = images[int(0.8*images_len):]\n    \n    def __len__(self):\n        return len(self.images)\n    \n    def __getitem__(self, idx):\n        image_path = self.images[idx]\n        data = Image.open(image_path)\n        if self.test:\n            return self.transform(data), image_path.split('.')[-2].split('/')[-1]\n        if 'cat' in image_path:\n            label = 0\n        else:\n            label = 1\n        return self.transform(data), label\n\ndata_transform = transforms.Compose([\n    transforms.Resize(256),\n    transforms.RandomCrop(224),\n    transforms.RandomHorizontalFlip(),\n    transforms.ToTensor()\n])\ntrain_set = DogCat(train_dir, transform=data_transform)\nvalid_set = DogCat(train_dir, train=False, transform=data_transform)\ntest_set = DogCat(test_dir, train=False, test=True, transform=data_transform)\ntrain_loader = DataLoader(train_set, batch_size=256, shuffle=True)\nvalid_loader = DataLoader(valid_set, batch_size=256, shuffle=True)\ntest_loader = DataLoader(test_set, batch_size=256, shuffle=True)\n# inputs, labels = iter(train_loader).next()\n# fig = plt.figure(figsize=(24, 16))\n# fig.tight_layout()\n# for idx in range(24):\n#     plt.subplot(4,6, idx+1)\n#     plt.title('dog' if labels[idx] == 1 else 'cat')\n#     plt.axis('off')\n#     # np.transpose()  returns the modified array.\n#     plt.imshow(np.transpose(inputs[idx], (1,2,0)))\n\ndevice = torch.device(\"cuda:0\" if torch.cuda.is_available() else \"cpu\")\nresnet_18 = torchvision.models.resnet18(num_classes=2)\nresnet_18.to(device)\ncriterion = nn.CrossEntropyLoss()\noptimizer = optim.SGD(resnet_18.parameters(), lr=0.001, momentum=0.9)\n\nmax_epoch = 64\nfor epoch in range(max_epoch):\n    for idx, (inputs,labels) in enumerate(train_loader):\n        inputs = inputs.to(device)\n        labels = labels.to(device)\n        optimizer.zero_grad()\n        outputs = resnet_18(inputs)\n        loss = criterion(outputs, labels)\n        loss.backward()\n        optimizer.step()\n        logger.debug('epoch: {}, batch: {}, loss: {}'.format(epoch+1, idx+1, loss.data))\ntorch.save(resnet_18.state_dict(), 'resnet_18.pth')\n\nresnet_18 = torchvision.models.resnet18(num_classes=2)\nresnet_18.load_state_dict(torch.load('resnet_18.pth'))\n\n# inputs, labels = iter(valid_loader).next()\n# outputs = resnet_18(inputs)\n# _, predicted = torch.max(outputs, 1)\n\n# fig = plt.figure(figsize=(24, 16))\n# fig.tight_layout()\n# for idx in range(24):\n#     plt.subplot(4,6, idx+1)\n#     plt.title('dog' if predicted[idx] == 1 else 'cat')\n#     plt.axis('off')\n#     # np.transpose()  returns the modified array.\n#     plt.imshow(np.transpose(inputs[idx], (1,2,0)))\n\n# test\nsubmission = pd.DataFrame(columns=('id', 'label'))\nfor inputs, image_id in test_loader:\n    outputs = resnet_18(inputs)\n    _, predicted = torch.max(outputs, 1)\n    for idx in range(predicted.size(0)):\n        submission = submission.append([{'id': image_id[idx], 'label': predicted[idx].data}])\nsubmission.to_csv('submission.csv', index=False)\n\n# valid\nwith torch.no_grad():\n    total = 0\n    correct = 0\n    for idx, data in enumerate(valid_loader):\n        images, labels = data\n        outputs = resnet_18(images)\n        _, predicted = torch.max(outputs.data, 1)\n        total += labels.size(0)\n        batch_correct = (predicted == labels).sum().item()\n        correct += batch_correct\n        logger.debug('batch:{}, accuracy:{}% '.format(idx+1, 100 * batch_correct / 256))\n    logger.debug('total:{}, correct:{}, total accuracy:{}% '.format(total, correct, correct*100 / total))","execution_count":null,"outputs":[]}],"metadata":{"kernelspec":{"language":"python","display_name":"Python 3","name":"python3"},"language_info":{"pygments_lexer":"ipython3","nbconvert_exporter":"python","version":"3.6.4","file_extension":".py","codemirror_mode":{"name":"ipython","version":3},"name":"python","mimetype":"text/x-python"}},"nbformat":4,"nbformat_minor":4}