python - RuntimeError: invalid argument 0: Sizes of tensors must match except in dimension 0

Question

when I am trying to create a DataLoader object with transforms I get this error. I am not sure why there are issues with tensor dimensions as shown here:

---------------------------------------------------------------------------
RuntimeError                              Traceback (most recent call last)
<ipython-input-3-e582b2a128f3> in <module>()
      8 classes = ('none', 'mild', 'moderate', 'severe', 'proliferative')
      9 
---> 10 for data, target in train_data_loader:
     11     for i in range(0, 4):
     12         im = data[i]

/share/software/user/open/py-pytorch/1.4.0_py36/lib/python3.6/site-packages/torch/utils/data/dataloader.py in __next__(self)
    343 
    344     def __next__(self):
--> 345         data = self._next_data()
    346         self._num_yielded += 1
    347         if self._dataset_kind == _DatasetKind.Iterable and 

/share/software/user/open/py-pytorch/1.4.0_py36/lib/python3.6/site-packages/torch/utils/data/dataloader.py in _next_data(self)
    854             else:
    855                 del self._task_info[idx]
--> 856                 return self._process_data(data)
    857 
    858     def _try_put_index(self):

/share/software/user/open/py-pytorch/1.4.0_py36/lib/python3.6/site-packages/torch/utils/data/dataloader.py in _process_data(self, data)
    879         self._try_put_index()
    880         if isinstance(data, ExceptionWrapper):
--> 881             data.reraise()
    882         return data
    883 

/share/software/user/open/py-pytorch/1.4.0_py36/lib/python3.6/site-packages/torch/_utils.py in reraise(self)
    392             # (https://bugs.python.org/issue2651), so we work around it.
    393             msg = KeyErrorMessage(msg)
--> 394         raise self.exc_type(msg)

RuntimeError: Caught RuntimeError in DataLoader worker process 0.
Original Traceback (most recent call last):
  File "/share/software/user/open/py-pytorch/1.4.0_py36/lib/python3.6/site-packages/torch/utils/data/_utils/worker.py", line 178, in _worker_loop
    data = fetcher.fetch(index)
  File "/share/software/user/open/py-pytorch/1.4.0_py36/lib/python3.6/site-packages/torch/utils/data/_utils/fetch.py", line 47, in fetch
    return self.collate_fn(data)
  File "/share/software/user/open/py-pytorch/1.4.0_py36/lib/python3.6/site-packages/torch/utils/data/_utils/collate.py", line 79, in default_collate
    return [default_collate(samples) for samples in transposed]
  File "/share/software/user/open/py-pytorch/1.4.0_py36/lib/python3.6/site-packages/torch/utils/data/_utils/collate.py", line 79, in <listcomp>
    return [default_collate(samples) for samples in transposed]
  File "/share/software/user/open/py-pytorch/1.4.0_py36/lib/python3.6/site-packages/torch/utils/data/_utils/collate.py", line 55, in default_collate
    return torch.stack(batch, 0, out=out)
RuntimeError: invalid argument 0: Sizes of tensors must match except in dimension 0. Got 634 and 440 in dimension 2 at /pytorch/aten/src/TH/generic/THTensor.cpp:612

Here's how you can reproduce the code. The error comes from creating a DataLoader object and applying the transforms to the images:

def check_cuda():
    is_cuda = False
    if torch.cuda.is_available():
        is_cuda = True
    return is_cuda

def scaleRadius(img, scale):
    x = img[int(img.shape[0] / 2), :, :].sum(1)
    r = (x > x.mean() / 10).sum() / 2
    s = scale * 1.0 / r
    return cv2.resize(img, (0,0), fx=s, fy=s)

class DRDataset(Dataset):
    def __init__(self, csv_path):
        # Transforms
        self.to_tensor = transforms.ToTensor()
        # Read the csv file
        self.data_info = pd.read_csv(csv_path, header=None)
        # First column contains the image paths
        self.image_arr = np.asarray(self.data_info.iloc[:, 0])
        # Second column is the labels
        self.label_arr = np.asarray(self.data_info.iloc[:, 1])
        # Calculate len
        self.data_len = len(self.data_info.index)

    def __getitem__(self, index):
        single_image_name = '/path/to/dataset' + self.image_arr[index] + '.jpeg'
        a = cv2.imread(single_image_name)
        
        scale = 300
        a = scaleRadius(a, scale)
        a = cv2.addWeighted(a, 4, cv2.GaussianBlur(a, (0,0), scale/30), -4, 128)
        b = np.zeros(a.shape)
        cv2.circle(b, (int(a.shape[1]/2), int(a.shape[0]/2)), int(scale*0.9), (1,1,1), -1, 8, 0)
        a = a*b + 128*(1-b)

        # Transform image to tensor
        img_as_tensor = self.to_tensor(a)

        # Get label(class) of the image based on the cropped pandas column
        single_image_label = self.label_arr[index]

        return (img_as_tensor, single_image_label)

    def __len__(self):
        return self.data_len
    
full_dataset = DRDataset('/path/to/dataset')

train_size = int(0.7 * len(full_dataset)) 
val_size = int(0.15 * len(full_dataset)) 
test_size = len(full_dataset) - train_size - val_size

train_dataset, val_dataset, test_dataset = torch.utils.data.random_split(full_dataset, 
                                                                         [train_size, val_size, test_size])

train_data_loader = torch.utils.data.DataLoader(train_dataset, batch_size=8,
                                          shuffle=True, num_workers=4, pin_memory=True)

valid_data_loader = torch.utils.data.DataLoader(val_dataset, batch_size=8,
                                          shuffle=True, num_workers=4, pin_memory=True)

for data, target in train_data_loader:
    for i in range(0, 4):  
        im = data[i]
        im = torch.squeeze(im)
        plt.imshow(np.transpose(im.numpy(), (1, 2, 0)), cmap='gray')
        plt.show()
        print(target[i] + ": " + classes[(int)(target[i])])
    break

What am I doing wrong?

--BJ

question from:https://stackoverflow.com/questions/65855641/runtimeerror-invalid-argument-0-sizes-of-tensors-must-match-except-in-dimensio

Answer 1

Waitting for answers