data.py

from config import *
from easydl import *
from collections import Counter
from torchvision.transforms.transforms import *
from torch.utils.data import DataLoader, WeightedRandomSampler

'''
assume classes across domains are the same.
[0 1 .......................................... N - 1]
|----common classes --||----source private classes --|
'''

source_classes = [i for i in range(args.data.dataset.n_total)]
target_classes = [i for i in range(args.data.dataset.n_share)]

train_transform = Compose([
    Resize(256),
    RandomCrop(224),
    RandomHorizontalFlip(),
    ToTensor()
])

test_transform = Compose([
    Resize(256),
    CenterCrop(224),
    ToTensor()
])

source_train_ds = FileListDataset(list_path=source_file, path_prefix=dataset.prefixes[args.data.dataset.source],
                            transform=train_transform, filter=(lambda x: x in source_classes))
source_test_ds = FileListDataset(list_path=source_file,path_prefix=dataset.prefixes[args.data.dataset.source],
                            transform=test_transform, filter=(lambda x: x in source_classes))
target_train_ds = FileListDataset(list_path=target_file, path_prefix=dataset.prefixes[args.data.dataset.target],
                            transform=train_transform, filter=(lambda x: x in target_classes))
target_test_ds = FileListDataset(list_path=target_file, path_prefix=dataset.prefixes[args.data.dataset.target],
                            transform=test_transform, filter=(lambda x: x in target_classes))

classes = source_train_ds.labels
freq = Counter(classes)
class_weight = {x : 1.0 / freq[x] if args.data.dataloader.class_balance else 1.0 for x in freq}

source_weights = [class_weight[x] for x in source_train_ds.labels]
sampler = WeightedRandomSampler(source_weights, len(source_train_ds.labels))

source_train_dl = DataLoader(dataset=source_train_ds, batch_size=args.data.dataloader.batch_size,
                             sampler=sampler, num_workers=args.data.dataloader.data_workers, drop_last=True)
source_test_dl = DataLoader(dataset=source_test_ds, batch_size=args.data.dataloader.batch_size, shuffle=False,
                             num_workers=1, drop_last=False)
target_train_dl = DataLoader(dataset=target_train_ds, batch_size=args.data.dataloader.batch_size,shuffle=True,
                             num_workers=args.data.dataloader.data_workers, drop_last=True)
target_test_dl = DataLoader(dataset=target_test_ds, batch_size=args.data.dataloader.batch_size, shuffle=False,
                             num_workers=1, drop_last=False)