CAMVID_PATH = Path('/bigguy/data', 'SegNet-Tutorial/CamVid') RESULTS_PATH = Path('.results/') WEIGHTS_PATH = Path('.weights/') RESULTS_PATH.mkdir(exist_ok=True) WEIGHTS_PATH.mkdir(exist_ok=True) batch_size = 2 normalize = transforms.Normalize(mean=camvid.mean, std=camvid.std) train_joint_transformer = transforms.Compose([ #joint_transforms.JointRandomCrop(224), # commented for fine-tuning joint_transforms.JointRandomHorizontalFlip() ]) train_dset = camvid.CamVid(CAMVID_PATH, 'train', joint_transform=train_joint_transformer, transform=transforms.Compose([ transforms.ToTensor(), normalize, ])) train_loader = torch.utils.data.DataLoader(train_dset, batch_size=batch_size, shuffle=True) val_dset = camvid.CamVid(CAMVID_PATH, 'val', joint_transform=None, transform=transforms.Compose( [transforms.ToTensor(), normalize])) val_loader = torch.utils.data.DataLoader(val_dset, batch_size=batch_size, shuffle=False)
import torchvision import torchvision.transforms as transforms from datasets import camvid import utils.imgs import utils.training as train_utils from datasets import joint_transforms from pathlib import Path from models import tiramisu CAMVID_PATH = Path('/home/jingwenlai/data', 'CamVid/CamVid') batch_size = 2 normalize = transforms.Normalize(mean=camvid.mean, std=camvid.std) test_dset = camvid.CamVid(CAMVID_PATH, 'test', joint_transform=None, transform=transforms.Compose( [transforms.ToTensor(), normalize])) test_loader = torch.utils.data.DataLoader(test_dset, batch_size=batch_size, shuffle=False) print("Test: %d" % len(test_loader.dataset.imgs)) model = tiramisu.FCDenseNet67(n_classes=12).cuda() model_weights = ".weights/latest.th" startEpoch = train_utils.load_weights(model, model_weights) print("load_weights, return epoch: ", startEpoch) train_utils.view_sample_predictions(model, test_loader, n=10)
CAMVID_PATH = "./CamVid/CamVid/" RESULTS_PATH = Path(".results/") WEIGHTS_PATH = Path(".weights/") RESULTS_PATH.mkdir(exist_ok=True) WEIGHTS_PATH.mkdir(exist_ok=True) batch_size = hyper["batch_size"] normalize = transforms.Normalize(mean=camvid.mean, std=camvid.std) train_joint_transformer = transforms.Compose([ # joint_transforms.JointRandomCrop(224), # commented for fine-tuning joint_transforms.JointRandomHorizontalFlip() ]) train_dset = camvid.CamVid( CAMVID_PATH, "train", joint_transform=train_joint_transformer, transform=transforms.Compose([transforms.ToTensor(), normalize]), ) train_loader = torch.utils.data.DataLoader(train_dset, batch_size=batch_size, shuffle=True) val_dset = camvid.CamVid( CAMVID_PATH, "val", joint_transform=None, transform=transforms.Compose([transforms.ToTensor(), normalize]), ) val_loader = torch.utils.data.DataLoader(val_dset, batch_size=batch_size,