making dataset with different channels

[zinnia1223]

Hello, I'm a newbie in Python programming and AI.
I have an x-ray data set for study. The problem is that... there are several images that have different channels.

I used these transforms.

train_transforms = Compose(
    [   
        LoadImage(image_only=True, ensure_channel_first = True),
        Resize(spatial_size = (224, 224)),
        NormalizeIntensity(),
        RandRotate(range_x=np.pi / 12, prob=0.5, keep_size=True),
        RandFlip(spatial_axis=0, prob=0.3),
        RandZoom(min_zoom=0.9, max_zoom=1.1, prob=0.5),
       
        EnsureType(),
    ]
)

class Dataset(torch.utils.data.Dataset):
    def __init__(self, image_files, labels, transforms):
        self.image_files = image_files
        self.labels = labels
        self.transforms = transforms

    def __len__(self):
        return len(self.image_files)

    def __getitem__(self, index):
        return self.transforms(self.image_files[index]), self.labels[index]

train_ds = Dataset(train_x, train_y, transforms = train_transforms)
train_loader = torch.utils.data.DataLoader(
    train_ds, batch_size = 5, shuffle = True, num_workers = 2)
val_ds = Dataset(val_x, val_y, val_transforms)
val_loader = torch.utils.data.DataLoader(
    val_ds, batch_size = 5, num_workers = 2
)
test_ds = Dataset(test_x, test_y, val_transforms)
test_loader = torch.utils.data.DataLoader(
    test_ds, batch_size = 5, num_workers = 2
)

for i in range(5):
    images, labels = train_ds[i]
    print(i, images.shape)

The result is...

0 torch.Size([3, 224, 224])
1 torch.Size([1, 224, 224])
2 torch.Size([1, 224, 224])
3 torch.Size([1, 224, 224])
4 torch.Size([1, 224, 224])

And I got an error while training a model as below:

RuntimeError: stack expects each tensor to be equal size, but got [1, 224, 224] at entry 0 and [3, 224, 224] at entry 2

Is there any way to change [3, 224, 224] to [1, 224, 224], by using transforms?

Thank you.