Data Augmentation for my dataset for image classification task

1

I want to perform a dog bred classifier task. The number of dogs per breed that I have is different and I want to perform data augmentation to make the frequency to be the same or to make the data balanced. Below is my code. How can I modify the below to not just transform my original dataset but transform and append it to existing dataset and eventually have the same number of classes

train_transforms = transforms.Compose([transforms.RandomRotation(30),
                                           transforms.RandomHorizontalFlip(),
                                           transforms.Resize(224),
                                           transforms.CenterCrop(224),
                                           transforms.ColorJitter(
                                                                    brightness=0.1*torch.randn(1),
                                                                    contrast=0.1*torch.randn(1),
                                                                    saturation=0.1*torch.randn(1),
                                                                    hue=0.1*torch.randn(1)),
                                           transforms.ToTensor(),
                                           transforms.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225])])

train_datasets = datasets.ImageFolder( os.path.join(data_dir,"train") , transform=train_transforms)
trainloader = torch.utils.data.DataLoader(train_datasets, batch_size=32, shuffle=True)
# dog_classes = trainloader.classes
data, target = iter(trainloader).next()