import os from argparse import ArgumentParser import numpy as np import torch import torch.nn.functional as F import torchvision.transforms as transforms from PIL import Image from torch.utils.data import DataLoader, Dataset import pytorch_lightning as pl from pl_examples.models.unet import UNet class KITTI(Dataset): """ Class for KITTI Semantic Segmentation Benchmark dataset Dataset link - http://www.cvlibs.net/datasets/kitti/eval_semseg.php?benchmark=semantics2015 There are 34 classes in the given labels. However, not all of them are useful for training (like railings on highways, road dividers, etc.). So, these useless classes (the pixel values of these classes) are stored in the `void_labels`. The useful classes are stored in the `valid_labels`. The `encode_segmap` function sets all pixels with any of the `void_labels` to `ignore_index` (250 by default). It also sets all of the valid pixels to the appropriate value between 0 and `len(valid_labels)` (since that is the number of valid classes), so it can be used properly by the loss function when comparing with the output. The `get_filenames` function retrieves the filenames of all images in the given `path` and saves the absolute path in a list. In the `get_item` function, images and masks are resized to the given `img_size`, masks are encoded using `encode_segmap`, and given `transform` (if any) are applied to the image only (mask does not usually require transforms, but they can be implemented in a similar way). """ def __init__( self, root_path, split='test', img_size=(1242, 376), void_labels=[0, 1, 2, 3, 4, 5, 6, 9, 10, 14, 15, 16, 18, 29, 30, -1], valid_labels=[7, 8, 11, 12, 13, 17, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 31, 32, 33], transform=None ): self.img_size = img_size self.void_labels = void_labels self.valid_labels = valid_labels self.ignore_index = 250 self.class_map = dict(zip(self.valid_labels, range(len(self.valid_labels)))) self.split = split self.root = root_path if self.split == 'train': self.img_path = os.path.join(self.root, 'training/image_2') self.mask_path = os.path.join(self.root, 'training/semantic') else: self.img_path = os.path.join(self.root, 'testing/image_2') self.mask_path = None self.transform = transform self.img_list = self.get_filenames(self.img_path) if self.split == 'train': self.mask_list = self.get_filenames(self.mask_path) else: self.mask_list = None def __len__(self): return len(self.img_list) def __getitem__(self, idx): img = Image.open(self.img_list[idx]) img = img.resize(self.img_size) img = np.array(img) if self.split == 'train': mask = Image.open(self.mask_list[idx]).convert('L') mask = mask.resize(self.img_size) mask = np.array(mask) mask = self.encode_segmap(mask) if self.transform: img = self.transform(img) if self.split == 'train': return img, mask else: return img def encode_segmap(self, mask): """ Sets void classes to zero so they won't be considered for training """ for voidc in self.void_labels: mask[mask == voidc] = self.ignore_index for validc in self.valid_labels: mask[mask == validc] = self.class_map[validc] return mask def get_filenames(self, path): """ Returns a list of absolute paths to images inside given `path` """ files_list = list() for filename in os.listdir(path): files_list.append(os.path.join(path, filename)) return files_list class SegModel(pl.LightningModule): """ Semantic Segmentation Module This is a basic semantic segmentation module implemented with Lightning. It uses CrossEntropyLoss as the default loss function. May be replaced with other loss functions as required. It is specific to KITTI dataset i.e. dataloaders are for KITTI and Normalize transform uses the mean and standard deviation of this dataset. It uses the FCN ResNet50 model as an example. Adam optimizer is used along with Cosine Annealing learning rate scheduler. """ def __init__(self, hparams): super().__init__() self.root_path = hparams.root self.batch_size = hparams.batch_size self.learning_rate = hparams.lr self.net = UNet(num_classes=19) self.transform = transforms.Compose([ transforms.ToTensor(), transforms.Normalize(mean=[0.35675976, 0.37380189, 0.3764753], std=[0.32064945, 0.32098866, 0.32325324]) ]) self.trainset = KITTI(self.root_path, split='train', transform=self.transform) self.testset = KITTI(self.root_path, split='test', transform=self.transform) def forward(self, x): return self.net(x) def training_step(self, batch, batch_nb): img, mask = batch img = img.float() mask = mask.long() out = self(img) loss_val = F.cross_entropy(out, mask, ignore_index=250) return {'loss': loss_val} def configure_optimizers(self): opt = torch.optim.Adam(self.net.parameters(), lr=self.learning_rate) sch = torch.optim.lr_scheduler.CosineAnnealingLR(opt, T_max=10) return [opt], [sch] def train_dataloader(self): return DataLoader(self.trainset, batch_size=self.batch_size, shuffle=True) def test_dataloader(self): return DataLoader(self.testset, batch_size=self.batch_size, shuffle=False) def main(hparams): # ------------------------ # 1 INIT LIGHTNING MODEL # ------------------------ model = SegModel(hparams) # ------------------------ # 2 INIT TRAINER # ------------------------ trainer = pl.Trainer( gpus=hparams.gpus ) # ------------------------ # 3 START TRAINING # ------------------------ trainer.fit(model) if __name__ == '__main__': parser = ArgumentParser() parser.add_argument("--root", type=str, help="path where dataset is stored") parser.add_argument("--gpus", type=int, help="number of available GPUs") parser.add_argument("--batch_size", type=int, default=4, help="size of the batches") parser.add_argument("--lr", type=float, default=0.001, help="adam: learning rate") hparams = parser.parse_args() main(hparams)