Using Data Augmentation to choose one class label

#1

In Pytorch
I want to use Data Augmentation to choose one class label.
Somebody tell me how to, please.
For example,
0 label class <- Use Data Augmentation
1 label class <- Not using

#2

You could add a condition in your Dataset's __getitem__:

class MyDataset(Dataset):
    def __init__(self, data, target, transform=None):
        ...
        self.transform = transform
        
    def __getitem__(self, index):
        x = self.data[index]
        y = self.target[index]
        if y == 0:
            x = self.transform(x)
        return x, y
    
    def __len__(self):
        return len(self.data)

You could also pass different transformations and apply them based on the current target label.

2 Likes
#3

Dear ptrblck

What beautiful code this is !
I see a lot.

Thanks for your help!

1 Like