Getting truth value of array with more than one element in ambigous for albuementation transform

Question

I am using albumentations for applying transform to a Pytorch model but getting this error and I m not getting any clue of what this error is about. Only thing I know is this is occuring due to transform which is being applied but not sure what is wrong with that.

ValueError: Traceback (most recent call last):
  File "/opt/conda/lib/python3.6/site-packages/torch/utils/data/_utils/worker.py", line 99, in _worker_loop
    samples = collate_fn([dataset[i] for i in batch_indices])
  File "/opt/conda/lib/python3.6/site-packages/torch/utils/data/_utils/worker.py", line 99, in <listcomp>
    samples = collate_fn([dataset[i] for i in batch_indices])
  File "<ipython-input-23-119ea6bc360e>", line 24, in __getitem__
    image = self.transform(image)
  File "/opt/conda/lib/python3.6/site-packages/albumentations/core/composition.py", line 164, in __call__
    need_to_run = force_apply or random.random() < self.p
ValueError: The truth value of an array with more than one element is ambiguous. Use a.any() or a.all()

This is are the snippets of code. Dataloader getitem() method:

        image = cv2.imread(p_path)
        image = cv2.cvtColor(image, cv2.COLOR_BGR2RGB)
        image = crop_image_from_gray(image)
        image = cv2.resize(image, (IMG_SIZE, IMG_SIZE))
        image = cv2.addWeighted ( image,4, cv2.GaussianBlur( image , (0,0) , 10) ,-4 ,128)
        print(image.shape)
        image = self.transform(image)

transforms applied :

val_transform = albumentations.Compose([
            Normalize(
                mean=[0.485, 0.456, 0.406],
                std=[0.229, 0.224, 0.225],
            ),
            ToTensor()
        ])

The class is called by:

valset       = MyDataset(val_df, transform = val_transform)

Ioannis Nasios · Answer 1 · 2019-09-02T11:20:00.223

From official albumentation documentation, you can apply transformation to image

from PIL import Image
import cv2
import numpy as np
from torch.utils.data import Dataset
from torchvision import transforms
from albumentations import Compose, RandomCrop, Normalize, HorizontalFlip, Resize
from albumentations.pytorch import ToTensor


class AlbumentationsDataset(Dataset):
    """__init__ and __len__ functions are the same as in TorchvisionDataset"""
    def __init__(self, file_paths, labels, transform=None):
        self.file_paths = file_paths
        self.labels = labels
        self.transform = transform

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

    def __getitem__(self, idx):
        label = self.labels[idx]
        file_path = self.file_paths[idx]

        # Read an image with OpenCV
        image = cv2.imread(file_path)

        # By default OpenCV uses BGR color space for color images,
        # so we need to convert the image to RGB color space.
        image = cv2.cvtColor(image, cv2.COLOR_BGR2RGB)

        image = crop_image_from_gray(image)
        image = cv2.resize(image, (IMG_SIZE, IMG_SIZE))
        image = cv2.addWeighted ( image,4, cv2.GaussianBlur( image , (0,0) , 10) ,-4 ,128)

        image = Img.fromarray(image, mode='RGB')  
        if self.transform:
            augmented = self.transform(image=np.array(image))
            image = augmented['image']

        image = np.transpose(image, (2, 0, 1))

        return image, label


albumentations_transform = Compose([

    Normalize(
        mean=[0.485, 0.456, 0.406],
        std=[0.229, 0.224, 0.225],
    ),
    ToTensor()
])


albumentations_dataset = AlbumentationsDataset(
    file_paths=['./images/image_1.jpg', './images/image_2.jpg', './images/image_3.jpg'],
    labels=[1, 2, 3],
    transform=albumentations_transform,
)

test_loader = DataLoader(dataset = albumentations_dataset, batch_size=4, drop_last=False, shuffle=False).

Ya albumentations are applied to numpy matrix images but can you tell me what issue is there with my piece of code. I mean after reading the image I have to do some image processing to be loaded to the model. Is that causing the issue or the way I m using albumentations — Sayooj Balakrishnan, Aug 30 '19 at 11:58
The way I m using albumentations is what causing the issue. Try my edited code above. — Ioannis Nasios, Aug 30 '19 at 13:06

Getting truth value of array with more than one element in ambigous for albuementation transform

1 Answers1