ValueError: Shapes must be equal rank, but are 1 and 0

Question

I am trying to use albumentations for my image recognition task but I have a little bit of problem while using this library. I have followed the documentation of albumentation in implementing this but I see the following error. ValueError: in user code:

File "/var/folders/ty/5qvxn8z53dg29rj28z94qmjr0000gn/T/ipykernel_1549/3096662177.py", line 3, in set_shapes  *
    label.set_shape([])

ValueError: Shapes must be equal rank, but are 1 and 0

Below are my code blocks

images = tf.keras.utils.image_dataset_from_directory(
    data_dir,
    labels='inferred',
    shuffle=True,
    image_size=(img_height, img_weight),
    )


transforms = Compose([
            Rotate(limit=40),
            RandomBrightness(limit=0.1),
            JpegCompression(quality_lower=85, quality_upper=100, p=0.5),
#             HueSaturationValue(hue_shift_limit=20, sat_shift_limit=30, val_shift_limit=20, p=0.5),
            RandomContrast(limit=0.2, p=0.5),
            HorizontalFlip(),
        ])


def aug_fn(image, img_size):
    data = {"image":image}
    aug_data = transforms(**data)
    aug_img = aug_data["image"]
    aug_img = tf.cast(aug_img, tf.float32)
    
    return aug_img

def process_data(image, label, img_size):
    aug_img = tf.numpy_function(func=aug_fn, inp=[image, img_size], Tout=tf.float32)
    return aug_img, label

aug_images= images.map(partial(process_data, img_size=112), num_parallel_calls=AUTOTUNE).prefetch(AUTOTUNE)
aug_images

def set_shapes(img, label, img_shape=(112,112,3)):
    img.set_shape(img_shape)
    label.set_shape([])
    return img, label

aug_images2= aug_images.map(set_shapes, num_parallel_calls=AUTOTUNE).prefetch(AUTOTUNE)
aug_images2

I get this error while trying to set the shapes back to the original form

ValueError: Shapes must be equal rank, but are 1 and 0

0 Answers0