Where to feed torchvision.transforms.Normalize()?

There was a transofrms field in DataLoader. I checked the manual it’s gone.
Where to feed transforms?
How to use it? It takes only mean and stddev.

from torchvision import transform

mean = [0.5071, 0.4867, 0.4408]
stdv = [0.2675, 0.2565, 0.2761]
train_transforms = transforms.Compose([
transforms.Normalize(mean=mean, std=stdv),

Solution is worker_init_fn= in DataLoader:

data = torch.utils.data.DataLoader((X_train, labels),worker_init_fn=transforms,batch_size=500)

Usually you would pass the transformation to a Dataset, not the DataLoader.
You could use e.g. torchvision.datasets.ImageFolder and pass it as transform=your_transform.
Alternatively, you could write your own Dataset and apply the transformation in the __getitem__ method.