{"cells":[{"metadata":{},"cell_type":"markdown","source":"# **Introduction**\n\nIt's been some weeks that the competition has started. The tiling method proposed by iafoss seems giving pretty good results on public LB. \nMajority of kernels use this method to fine-tune a CNN (resnet and its variants are often used) and predict isup grade with classification or regression.\n\nIn this kernel I've tackled a different way of classifying biopsy tissues.\nIndeed CNNs assume that inputs are independant of each other while RNNs assume that there is an interaction between the input sequences. The CNN could be used to extract tile features and the RNN could consider them as a sequence. The aim would be to extract the long-term dependencies of the features sequence to improve the classification accuracy. \n\n![image.png](attachment:image.png)\n\n1. This is a starter kernel. Unfortunately I don't have my own hardware so I cannot do all the tests that I want because of GPU quota. I hope this kernel will provide ideas for further improvements !\n2. I've \"frozen\" the backbone but one could train some epochs with that configuration and pursue training with the entire model \n3. One could also train the LSTM in pararell of the CNN and fuse features from both models, use some kind of attention layer to keep relevant features and perform classication.\n\n\nI've used a similar training framework that https://www.kaggle.com/yasufuminakama/panda-se-resnext50-classification-baseline + iafoss tiling method https://www.kaggle.com/iafoss/panda-concat-tile-pooling-starter-0-79-lb ","attachments":{"image.png":{"image/png":"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"}},"execution_count":null},{"metadata":{},"cell_type":"markdown","source":"# Libraries","execution_count":null},{"metadata":{"_uuid":"8f2839f25d086af736a60e9eeb907d3b93b6e0e5","_cell_guid":"b1076dfc-b9ad-4769-8c92-a6c4dae69d19","trusted":true},"cell_type":"code","source":"import os\nimport gc\nimport random\nimport sys\nimport time\nfrom contextlib import contextmanager\nfrom pathlib import Path\nfrom collections import defaultdict, Counter\nimport numpy as np\nimport pandas as pd\nfrom tqdm.notebook import tqdm\nimport matplotlib.pyplot as plt\nimport skimage.io\nimport cv2\nimport PIL\nfrom albumentations.pytorch import ToTensorV2\nfrom albumentations import Compose, Normalize, HorizontalFlip, VerticalFlip, Resize, \\\nRandomRotate90, OneOf, RandomContrast, RandomGamma, RandomBrightness, ShiftScaleRotate\nimport torch\nfrom skimage.transform import AffineTransform, warp\nimport warnings\nwarnings.filterwarnings(\"ignore\")\nimport torchvision\nimport torch.nn.functional as F\nfrom torch.utils.data.dataloader import DataLoader\nimport torch.nn as nn\nfrom torch.optim.lr_scheduler import ReduceLROnPlateau, OneCycleLR\nfrom sklearn.model_selection import StratifiedKFold\nfrom sklearn.metrics import cohen_kappa_score, confusion_matrix, accuracy_score\nimport seaborn as sn\nfrom functools import partial\n\n# Install pre-trained models\nsys.path.insert(0, '../input/pytorch-pretrained-models/semi-supervised-ImageNet1K-models-master/semi-supervised-ImageNet1K-models-master/')\nfrom hubconf import *\n\n#Define paths\nBASE_PATH = '/kaggle/input'\nTRAIN_IMG_DIR = f'{BASE_PATH}/panda-128x128x20/kaggle/train_images/'\ntrain = pd.read_csv(f'{BASE_PATH}/prostate-cancer-grade-assessment/train.csv').set_index('image_id')\n\n# Global variables\nMODEL_NAME = 'resnext50_32x4d_ssl' \n\nLOSS = 'CE' ","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"# Config","execution_count":null},{"metadata":{"trusted":true},"cell_type":"code","source":"class CFG:\n    debug=False\n    lr=1e-3\n    batch_size=16\n    epochs=5\n    onecyclepolicy=False\n    seed=35\n    target_col='isup_grade'\n    n_fold=4\n    \nclass CFG_MODEL:\n    n_out = 6 # number of classes\n    lstm_layer = 2 # number of LSTM layers\n    lstm_hidden_sz = 512 # Number of features in hidden state","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"files = sorted(set([p[:32] for p in os.listdir(TRAIN_IMG_DIR)]))\ntrain = train.loc[files]\ntrain = train.reset_index()","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"# Utils","execution_count":null},{"metadata":{"_uuid":"d629ff2d2480ee46fbb7e2d37f6b5fab8052498a","_cell_guid":"79c7e3d0-c299-4dcb-8224-4455121ee9b0","trusted":true},"cell_type":"code","source":"def seed_everything(seed=99):\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    torch.backends.cudnn.enabled = True\n    \n@contextmanager\ndef timer(name):\n    t0 = time.time()\n    LOGGER.info(f'[{name}] start')\n    yield\n    LOGGER.info(f'[{name}] done in {time.time() - t0:.0f} s.')\n\n    \ndef init_logger(log_file='train.log'):\n    from logging import getLogger, DEBUG, FileHandler,  Formatter,  StreamHandler\n    \n    log_format = '%(asctime)s %(levelname)s %(message)s'\n    \n    stream_handler = StreamHandler()\n    stream_handler.setLevel(DEBUG)\n    stream_handler.setFormatter(Formatter(log_format))\n    \n    file_handler = FileHandler(log_file)\n    file_handler.setFormatter(Formatter(log_format))\n    \n    logger = getLogger('PANDA')\n    logger.setLevel(DEBUG)\n    logger.addHandler(stream_handler)\n    logger.addHandler(file_handler)\n    \n    return logger\n\nLOG_FILE = 'train.log'\nLOGGER = init_logger(LOG_FILE)","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"# Dataset","execution_count":null},{"metadata":{"trusted":true},"cell_type":"code","source":"class PandaDatasetInt:\n    def __init__(self, df, N, sz, transform=None):\n        self.image_ids=df.image_id.values\n        self.isup_grade = df.isup_grade.values\n        self.data_provider = df.data_provider.values\n        self.transform=transform\n        self.df=df\n        self.tile_sz = sz\n        self.tile_nb = N\n    def __len__(self):\n        return len(self.image_ids)\n    \n    def __getitem__(self, index):\n        fn_base = f'{TRAIN_IMG_DIR}{self.image_ids[index]}'\n        img = PIL.Image.open(f\"{fn_base}.png\").convert('RGB')\n        img = np.array(img)\n        img = img.reshape(img.shape[0] // self.tile_sz, self.tile_sz, img.shape[1] // self.tile_sz, self.tile_sz, 3)\n        img = img.transpose(0, 2, 1, 3, 4).reshape(-1, self.tile_sz, self.tile_sz, 3)  # Nxszxszx3\n        img = img.astype(np.single)\n        \n        if self.transform:\n            for i in range(self.tile_nb):\n                aug = self.transform(image=img[i])\n                img[i] = aug['image'].permute(1,2,0)\n                \n        label = self.isup_grade[index]\n        provider = self.data_provider[index]\n        return  torch.tensor(img.transpose(0, 3, 1, 2)), torch.tensor(label), provider","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"# Visualize","execution_count":null},{"metadata":{"trusted":true},"cell_type":"code","source":"# Display batch\ntfm = Compose([\n        ToTensorV2(),\n    ])\n\n\ndataset = PandaDatasetInt(train, N=20, sz=128, transform=tfm)\nloader = DataLoader(dataset, batch_size=CFG.batch_size, shuffle=False)\n\nimages, label, provider = next(iter(loader))\nfor i in range(8):\n    grid = torchvision.utils.make_grid(images[i].int(), nrow=20)\n    plt.figure(figsize=(30,30))\n    plt.imshow(grid.permute(1,2,0))\n    plt.title(f'isup grade = {label[i].item()} | provider : {provider[i]}')","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"# Define model","execution_count":null},{"metadata":{"trusted":true},"cell_type":"code","source":"def _resnext(block, layers, pretrained, progress, **kwargs):\n    model = ResNet(block, layers, **kwargs)\n\n    return model\n\nclass Model(nn.Module):\n    def __init__(self, n = 6, bs = 16, nb_layer = 1, hidden = 512, device = 'cuda'):\n        super().__init__()\n        self.out_dim = n\n        self.hidden_sz = hidden\n        self.layer_dim = nb_layer\n        self.device = device\n        \n        # ResNext\n        m = _resnext(Bottleneck, [3, 4, 6, 3], False, progress=False, groups=32, width_per_group=4)\n        m.load_state_dict(torch.load('../input/pytorch-pretrained-models/semi_supervised_resnext50_32x4-ddb3e555.pth'))\n        self.backbone = nn.Sequential(*list(m.children())[:-1])\n        nc = list(m.children())[-1].in_features  # 2048\n        \n        # LSTM\n        self.lstm = nn.LSTM(nc, self.hidden_sz, self.layer_dim, dropout=0.5, batch_first=True, bidirectional=False)\n        \n        # FC              \n        self.fc = nn.Linear(self.hidden_sz, self.out_dim)\n\n    def forward(self, x):\n        # CNN\n        n = x.shape[1]\n        x = x.view(-1, x.shape[2], x.shape[3], x.shape[4]) # => x: bs*N x 3 x 128 x 128\n        x = self.backbone(x) # => x: bs*N x C x 1 x 1\n        x = x.view(-1, n, x.shape[1], x.shape[2], x.shape[3]) # => x: bs x N x C x 1 x 1\n        x = x.view(x.shape[0], x.shape[1], -1) # => x: bs x N x C\n        # Shuffle the tile features\n        row_idxs = list(range(x.shape[1]))\n        random.shuffle(row_idxs)\n        x = x[:, torch.tensor(row_idxs)]\n        \n        # LSTM\n        # Set initial hidden states\n        h0 = torch.zeros(self.layer_dim, x.shape[0], self.hidden_sz).to(self.device)\n        c0 = torch.zeros(self.layer_dim, x.shape[0], self.hidden_sz).to(self.device)\n        out,  h  = self.lstm(x, (h0, c0)) # => x: bs x hidden\n        \n        # FC\n        x = self.fc(out[:, -1, :]) # => x: bs x 6\n        return x","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"if CFG.debug:\n    folds = train.sample(n=50, random_state=CFG.seed).reset_index(drop=True).copy()\nelse:\n    folds = train.copy()\n\ntrain_labels = folds[CFG.target_col].values\nkf = StratifiedKFold(n_splits=CFG.n_fold, shuffle=True, random_state=CFG.seed)\nfor fold, (train_index, val_index) in enumerate(kf.split(folds.values, train_labels)):\n    folds.loc[val_index, 'fold'] = int(fold)\nfolds['fold'] = folds['fold'].astype(int)\nfolds.to_csv('folds.csv', index=None)\nfolds.head()","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"def freeze(m):\n    for p in m.parameters():\n        p.requires_grad = False\n\ndef unfreeze(m):\n    for p in m.parameters():\n        p.requires_grad = True","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"def train_fn(fold):\n    device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')\n    print(f\"### fold: {fold} ###\")\n    \n    lossTrain = []\n    stepTrain = []\n    lossVal = []\n    stepVal = []\n    \n    step = 0\n        \n    trn_idx = folds[folds['fold'] != fold].index\n    val_idx = folds[folds['fold'] == fold].index\n    \n    transformTrain = Compose([\n        OneOf([RandomBrightness(limit=0.15), RandomContrast(limit=0.3), RandomGamma()], p=0.25),\n        HorizontalFlip(p=0.5),\n        VerticalFlip(p=0.5),\n        ShiftScaleRotate(shift_limit=0.1, scale_limit=0.1, rotate_limit=20, p=0.3),\n        Normalize(\n            mean=[0.485, 0.456, 0.406],\n            std=[0.229, 0.224, 0.225],\n        ),\n        ToTensorV2(),\n    ])\n\n    transformValid = Compose([\n            Normalize(\n                mean=[0.485, 0.456, 0.406],\n                std=[0.229, 0.224, 0.225],\n            ),\n            ToTensorV2(),\n        ])\n                        \n    train_dataset = PandaDatasetInt(folds.loc[trn_idx].reset_index(drop=True), N=20, sz=128, transform=transformTrain)\n    valid_dataset = PandaDatasetInt(folds.loc[val_idx].reset_index(drop=True), N=20, sz=128, transform=transformValid)\n    \n    train_loader = DataLoader(train_dataset, batch_size=CFG.batch_size, num_workers=4)\n    valid_loader = DataLoader(valid_dataset, batch_size=CFG.batch_size, num_workers=4)\n    \n    model = Model(n=CFG_MODEL.n_out, bs=CFG.batch_size, nb_layer=CFG_MODEL.lstm_layer, hidden=CFG_MODEL.lstm_hidden_sz, device=device)\n    model.to(device)\n    \n    optimizer = torch.optim.Adam(model.parameters(), lr=CFG.lr, amsgrad=False)\n    if CFG.onecyclepolicy == True:\n        scheduler = torch.optim.lr_scheduler.OneCycleLR(optimizer, max_lr=CFG.lr, div_factor=100, pct_start=0.0, steps_per_epoch=len(train_loader), epochs=CFG.epochs)\n    else:\n        scheduler = ReduceLROnPlateau(optimizer, 'min', factor=0.5, patience=2, verbose=True, eps=1e-6)\n    \n    if LOSS == 'CE':\n        criterion = nn.CrossEntropyLoss()\n    elif LOSS == 'LabelSmoothingCE':\n        criterion = label_smoothing_criterion()\n        \n    best_score = -100\n    best_loss = np.inf\n    \n    # Train only the RNN and the FC\n    freeze(model.backbone)\n    \n    for epoch in range(CFG.epochs):\n        start_time = time.time()\n        \n        # Unfreeze backbone\n        #if epoch == 5:\n        #    unfreeze(model.backbone)\n            \n        # --------- Training loop ----------\n        model.train()\n        avg_loss = 0.\n\n        optimizer.zero_grad()\n        tk0 = tqdm(enumerate(train_loader), total=len(train_loader))\n\n        for i, (images, labels, _) in tk0:\n\n            images = images.to(device)\n            labels = labels.to(device)\n            \n            y_preds = model(images)\n            loss = criterion(y_preds, labels)\n            \n            loss.backward()\n            optimizer.step()\n            optimizer.zero_grad()\n            \n            lossTrain.append(loss.item())\n            stepTrain.append(step)\n            \n            if CFG.onecyclepolicy == True:\n                scheduler.step()\n\n            avg_loss += loss.item() / len(train_loader)\n            step += 1\n            \n        # ---------- Validation loop --------\n        model.eval()\n        avg_val_loss = 0.\n        preds = []\n        valid_labels = []\n        tk1 = tqdm(enumerate(valid_loader), total=len(valid_loader))\n\n        for i, (images, labels, _) in tk1:\n            \n            images = images.to(device)\n            labels = labels.to(device)\n            \n            with torch.no_grad():\n                y_preds = model(images)\n            \n            preds.append(y_preds.to('cpu').numpy().argmax(1))\n            valid_labels.append(labels.to('cpu').numpy())\n\n            loss = criterion(y_preds, labels)\n            avg_val_loss += loss.item() / len(valid_loader)\n            \n        lossVal.append(avg_val_loss)\n        stepVal.append(step)\n            \n        if CFG.onecyclepolicy == False:\n            scheduler.step(avg_val_loss)\n            \n        preds = np.concatenate(preds)\n        valid_labels = np.concatenate(valid_labels)\n        \n        LOGGER.debug(f'Counter preds: {Counter(preds)}')\n        score = cohen_kappa_score(valid_labels, preds, weights='quadratic')\n        \n        if epoch == (CFG.epochs - 1):\n            print(confusion_matrix(valid_labels, preds))\n\n        elapsed = time.time() - start_time\n        \n        LOGGER.debug(f'  Epoch {epoch+1} - avg_train_loss: {avg_loss:.4f}  avg_val_loss: {avg_val_loss:.4f}  time: {elapsed:.0f}s')\n        LOGGER.debug(f'  Epoch {epoch+1} - QWK: {score}')\n        \n        if score>best_score:\n            best_score = score\n            LOGGER.debug(f'  Epoch {epoch+1} - Save Best Score: {best_score:.4f} Model')\n            torch.save(model.state_dict(), f'fold{fold}_se_resnext50.pth')\n            \n    # Plot losses \n    plt.figure(figsize=(26,6))\n    plt.subplot(1, 2, 1)\n    plt.plot(stepTrain, lossTrain, label=\"training loss\")\n    plt.plot(stepVal, lossVal, label = \"validation loss\")\n    plt.title('Loss')\n    plt.xlabel('step')\n    plt.legend(loc='center left')\n    plt.tight_layout()\n    plt.show()","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"train_fn(0)","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}