{"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 20GB 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":"package_paths = '../input/pytorch-image-models/pytorch-image-models-master'\nimport sys\nsys.path.append(package_paths)\nimport os\nimport numpy as np\nimport pandas as pd\nimport torch\nfrom tqdm import tqdm\nimport cv2\nimport torch.nn as nn\nfrom torch.utils.data import Dataset,DataLoader\nimport timm\nimport random\nimport time\nfrom torch.cuda.amp import autocast, GradScaler","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"model_arch = 'tf_efficientnet_b4_ns'\ntrain_rate = 0.8\nvalid_rate = 0.2\ndata_root = '../input/cassava-leaf-disease-classification/train_images'\nimg_size = 512\nbatch_size = 16\nnum_workers = 4\nseed = 719\nepochs = 6\nlr = 1e-4\nweight_decay = 1e-6\ndevice = torch.device(\"cuda:0\" if torch.cuda.is_available() else \"cpu\")\nCFG = {\n    'verbose_step': 1,\n    'accum_iter': 2,  # suppoprt to do batch accumulation for backprop with effectively larger batch size\n    \n    }\n\nfrom albumentations import (\n    HorizontalFlip, VerticalFlip, IAAPerspective, ShiftScaleRotate, CLAHE, RandomRotate90,\n    Transpose, ShiftScaleRotate, Blur, OpticalDistortion, GridDistortion, HueSaturationValue,\n    IAAAdditiveGaussianNoise, GaussNoise, MotionBlur, MedianBlur, IAAPiecewiseAffine, RandomResizedCrop,\n    IAASharpen, IAAEmboss, RandomBrightnessContrast, Flip, OneOf, Compose, Normalize, Cutout, CoarseDropout, ShiftScaleRotate, CenterCrop, Resize\n)\nfrom albumentations.pytorch import ToTensorV2\ntrain_transforms = Compose([\n            RandomResizedCrop(img_size, img_size),\n            Transpose(p=0.5),\n            HorizontalFlip(p=0.5),\n            VerticalFlip(p=0.5),\n            ShiftScaleRotate(p=0.5),\n            HueSaturationValue(hue_shift_limit=0.2, sat_shift_limit=0.2, val_shift_limit=0.2, p=0.5),\n            RandomBrightnessContrast(brightness_limit=(-0.1,0.1), contrast_limit=(-0.1, 0.1), p=0.5),\n            Normalize(mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225], max_pixel_value=255.0, p=1.0),\n            CoarseDropout(p=0.5),\n            Cutout(p=0.5),\n            ToTensorV2(p=1.0),\n        ], p=1.)\n\nvalid_transforms = Compose([\n            CenterCrop(img_size, img_size, p=1.),\n            Resize(img_size, img_size),\n            Normalize(mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225], max_pixel_value=255.0, p=1.0),\n            ToTensorV2(p=1.0),\n        ], p=1.)","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"def seed_everything(seed):\n    random.seed(seed)\n    os.environ['PYTHONHASHSEED'] = str(seed)\n    np.random.seed(seed)\n    torch.manual_seed(seed)\n    torch.cuda.manual_seed(seed)\n    torch.backends.cudnn.deterministic = True\n    torch.backends.cudnn.benchmark = True\n\ndef get_image(path):\n    img_bgr = cv2.imread(path)\n    img_rgb = img_bgr[:,:,::-1]\n    return img_rgb","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"train_data = pd.read_csv('../input/cassava-leaf-disease-classification/train.csv')","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"imgtype = []","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"class CassavaDataset(Dataset):\n    def __init__(self,train,data_root,transforms):\n        super(CassavaDataset, self).__init__()\n        self.train = train\n        self.data_root = data_root\n        self.transforms = transforms\n    \n    def __len__(self):\n        return self.train.shape[0]\n    \n    def __getitem__(self, index):\n        target = self.train.label[index]\n        img = get_image('{}/{}'.format(self.data_root, self.train.loc[index]['image_id']))\n        if self.transforms:\n            img = self.transforms(image = img)['image']\n        imgtype.append(type(img))\n        return img, target\n        ","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"class CassavaImgClassifier(nn.Module):\n    def __init__(self, model_arch, n_class, pretrained=False):\n        super(CassavaImgClassifier, self).__init__()\n        self.model = timm.create_model(model_arch,pretrained=pretrained)\n        n_features = self.model.classifier.in_features\n        self.model.classifier = nn.Linear(n_features, n_class)\n        \n    def forward(self, x):\n        x = self.model(x)\n        return x","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"train = train_data[:int(train_data.shape[0]*train_rate)]\nvalid = train_data.iloc[int(train_data.shape[0]*train_rate):].reset_index(drop=True)\n\ntrain_DS = CassavaDataset(train,data_root,train_transforms)\nvalid_DS = CassavaDataset(valid,data_root,valid_transforms)\n\ntrain_loader = torch.utils.data.DataLoader(train_DS,\n                                           batch_size = batch_size,\n                                           shuffle = True,\n                                           num_workers = num_workers)\nvalid_loader = torch.utils.data.DataLoader(valid_DS,\n                                           batch_size = batch_size,\n                                           shuffle = True,\n                                           num_workers = num_workers)","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"if __name__ == '__main__':\n    seed_everything(seed)\n    \n    model = CassavaImgClassifier(model_arch,\n                                     train.label.nunique(), \n                                     pretrained=True).to(device)\n    \n    scaler = GradScaler()\n    optimizer = torch.optim.Adam(model.parameters(),\n                                     lr=lr, weight_decay=weight_decay)\n        \n    scheduler = torch.optim.lr_scheduler.CosineAnnealingWarmRestarts(\n            optimizer, T_0=10, T_mult=1, eta_min=1e-6, last_epoch=-1)\n        \n    loss_fn = nn.CrossEntropyLoss().to(device)\n        \n    for epoch in range(epochs):\n        # 训练\n        model.train()\n        running_loss = None\n        print('Training with {} started'.format(epoch))\n        print(len(train), len(valid))\n        \n        t = time.time()\n        pbar = tqdm(enumerate(train_loader), total=len(train_loader))\n        for step, (imgs, image_labels) in pbar:\n            imgs = imgs.to(device).float()\n            image_labels = image_labels.to(device).long()\n            \n            with autocast():\n                image_preds = model(imgs)\n                loss = loss_fn(image_preds, image_labels)\n                scaler.scale(loss).backward()\n                \n                if running_loss is None:\n                    running_loss = loss.item()\n                else:\n                    running_loss = running_loss * .99 + loss.item() * .01\n                \n                if ((step + 1) %  CFG['accum_iter'] == 0) or ((step + 1) == len(train_loader)):\n                    # may unscale_ here if desired (e.g., to allow clipping unscaled gradients)\n\n                    scaler.step(optimizer)\n                    scaler.update()\n                    optimizer.zero_grad() \n                \n\n                if ((step + 1) % CFG['verbose_step'] == 0) or ((step + 1) == len(train_loader)):\n                    description = f'epoch {epoch} loss: {running_loss:.4f}'\n                    pbar.set_description(description)\n        \n            if scheduler is not None and step%300==0:\n                scheduler.step()    \n        \n        # 验证\n        with torch.no_grad():\n            model.eval()\n            \n            t=time.time()\n            loss_sum = 0\n            sample_num = 0\n            image_preds_all = []\n            image_targets_all = []\n            \n            pbar = tqdm(enumerate(valid_loader), total=len(valid_loader))\n            for step, (imgs, image_labels) in pbar:\n                imgs = imgs.to(device).float()\n                image_labels = image_labels.to(device).long()\n\n                image_preds = model(imgs)   #output = model(input)\n                #print(image_preds.shape, exam_pred.shape)\n                image_preds_all += [torch.argmax(image_preds, 1).detach().cpu().numpy()]\n                image_targets_all += [image_labels.detach().cpu().numpy()]\n\n                loss = loss_fn(image_preds, image_labels)\n\n                loss_sum += loss.item()*image_labels.shape[0]\n                sample_num += image_labels.shape[0]  \n\n                if ((step + 1) % CFG['verbose_step'] == 0) or ((step + 1) == len(val_loader)):\n                    description = f'epoch {epoch} loss: {loss_sum/sample_num:.4f}'\n                    pbar.set_description(description)\n\n            image_preds_all = np.concatenate(image_preds_all)\n            image_targets_all = np.concatenate(image_targets_all)\n            print('validation multi-class accuracy = {:.4f}'.format((image_preds_all==image_targets_all).mean()))\n\n\n        torch.save(model.state_dict(),'{}_epoch_{}'.format(model_arch, epoch))\n\n    \n    del model, optimizer, train_loader, val_loader, scaler, scheduler\n    torch.cuda.empty_cache()\n    ","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"imgtype","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}