batch load with different length edge_attr

[xyqbest]

I have a dataset like this,

dataset[0]
Out[1]: Data(x=[396, 6], x2=[396, 23], x3=[396, 31], x4=[396, 31], y=[396, 2], el=[2, 63497], weight_sum=[63497], weight_mean=[63497], date_time=[1])

dataset[1]
Out[1]: Data(x=[396, 6], x2=[396, 23], x3=[396, 31], x4=[396, 31], y=[396, 2], el=[2, 57511], weight_sum=[57511], weight_mean=[57511], date_time=[1])

when I use this

from myDatasets import myDataset1
from torch_geometric.loader import DataLoader
dataset = myDataset1("./data")
loader = DataLoader(dataset, batch_size=10)
tmp = next(iter(loader))

to load data.
there are some errors, and I do not know how to solve it. Must each graph have same edges?

Traceback (most recent call last):
  File "D:\xxx\pythonProject\phdThesis\del_debug.py", line 8, in <module>
    tmp = next(iter(loader))
  File "D:\ProgramData\Anaconda3\envs\phdThesis\lib\site-packages\torch\utils\data\dataloader.py", line 521, in __next__
    data = self._next_data()
  File "D:\ProgramData\Anaconda3\envs\phdThesis\lib\site-packages\torch\utils\data\dataloader.py", line 561, in _next_data
    data = self._dataset_fetcher.fetch(index)  # may raise StopIteration
  File "D:\ProgramData\Anaconda3\envs\phdThesis\lib\site-packages\torch\utils\data\_utils\fetch.py", line 52, in fetch
    return self.collate_fn(data)
  File "D:\ProgramData\Anaconda3\envs\phdThesis\lib\site-packages\torch_geometric\loader\dataloader.py", line 39, in __call__
    return self.collate(batch)
  File "D:\ProgramData\Anaconda3\envs\phdThesis\lib\site-packages\torch_geometric\loader\dataloader.py", line 19, in collate
    return Batch.from_data_list(batch, self.follow_batch,
  File "D:\ProgramData\Anaconda3\envs\phdThesis\lib\site-packages\torch_geometric\data\batch.py", line 69, in from_data_list
    batch, slice_dict, inc_dict = collate(
  File "D:\ProgramData\Anaconda3\envs\phdThesis\lib\site-packages\torch_geometric\data\collate.py", line 85, in collate
    value, slices, incs = _collate(attr, values, data_list, stores,
  File "D:\ProgramData\Anaconda3\envs\phdThesis\lib\site-packages\torch_geometric\data\collate.py", line 162, in _collate
    value = torch.cat(values, dim=cat_dim or 0)
RuntimeError: Sizes of tensors must match except in dimension 0. Expected size 63497 but got size 57511 for tensor number 1 in the list.

BTW, it seems that I have to use .view to reshape the data from DataLoader if I want to use torch_geometric.loader.DataLoader to load image data?

[rusty1s]

Can you try to rename el to edge_index? Otherwise, it will try to concatenate the tensors in the first dimension :)

[xyqbest]

Thank you very much, It works