TWO RESIZING DATA AUGMENTATION

[EliiiFlowers]

HI EVERYONE, I WANT TO KNOW IF IS POSSIBLE DOING 2 RESIZING DATA AUGMENTATION, MY ORIGINAL IMAGE IS 512X512, BUT I WANT TO MAKE A FIRST RESIZE: 512X512 TO 256X256 AND OTHER FROM 512X512 TO 1024X1024. THIS IS MY CODE:
I WANT TO KNOW IF THIS IS CORRECT OR I'M MISSING SOMETHING

def custom_mapper(dataset_dict):
    # Implement a mapper, similar to the default DatasetMapper, but with your own customizations
    dataset_dict = copy.deepcopy(dataset_dict)  # it will be modified by code below
    image = utils.read_image(dataset_dict["file_name"], format="BGR")
    transform_list = [T.Resize((256,256)),
                      T.Resize((1024,1024)), 
                      T.RandomRotation(angle=[90, 90]),
                      T.RandomFlip(prob=0.5, horizontal=False, vertical=True),
                      T.RandomFlip(prob=0.5, horizontal=True, vertical=False), 
                    
                      ]
    image, transforms = T.apply_transform_gens(transform_list, image)
    dataset_dict["image"] = torch.as_tensor(image.transpose(2, 0, 1).astype("float32"))

    annos = [
        utils.transform_instance_annotations(obj, transforms, image.shape[:2])
        for obj in dataset_dict.pop("annotations")
        if obj.get("iscrowd", 0) == 0
    ]
    instances = utils.annotations_to_instances(annos, image.shape[:2])
    dataset_dict["instances"] = utils.filter_empty_instances(instances)
    return dataset_dict

[ExtReMLapin]

I could be mis-understanding something but as far as I know, on each minibatch creation, it runs all the images from the minibatch thru your transform list, which mean it doesn't create an image added to your base image, it replaces your image by the one modified with the transformList.

So it's not possible to have the same image twice (even if modified) in the same minibatch.