Data transformation

From now on, we recommend using our discussion forum (https://github.com/rusty1s/pytorch_geometric/discussions) for general questions.

❓ Questions & Help

I have the measured data saved in txt format and I want to use it to train the model. After processing, my Data object looks like following: Data(edge_attr=[1325], edge_index=[2, 1325], test_mask=[0], train_mask=[0], val_mask=[0], x=[919, 20], y=[919]) and this is directed graph:

print(graph_raw.is_undirected())
tensor(False)

But I want to make it undirected, so I apply transformation:

import torch_geometric.transforms as T
transf_to_und = T.ToUndirected()
transf_to_und(graph_raw)

out: Data(edge_attr=[1325], edge_index=[2, 2650], test_mask=[0], train_mask=[0], val_mask=[0], x=[919, 20], y=[919])

but when I check

print(graph_raw.is_undirected())

I receive an assertion error

AssertionError                            Traceback (most recent call last)
<ipython-input-131-0aa7e60ed235> in <module>
      2 transf_to_und = T.ToUndirected()
      3 transf_to_und(graph_raw)
----> 4 print(graph_raw.is_undirected())

~\anaconda3\lib\site-packages\torch_geometric\data\data.py in is_undirected(self)
    295     def is_undirected(self):
    296         r"""Returns :obj:`True`, if graph edges are undirected."""
--> 297         return is_undirected(self.edge_index, self.edge_attr, self.num_nodes)
    298 
    299     def is_directed(self):

~\anaconda3\lib\site-packages\torch_geometric\utils\undirected.py in is_undirected(edge_index, edge_attr, num_nodes)
     19     """
     20     num_nodes = maybe_num_nodes(edge_index, num_nodes)
---> 21     edge_index, edge_attr = coalesce(edge_index, edge_attr, num_nodes,
     22                                      num_nodes)
     23 

~\anaconda3\lib\site-packages\torch_sparse\coalesce.py in coalesce(index, value, m, n, op)
     20     """
     21 
---> 22     storage = SparseStorage(row=index[0], col=index[1], value=value,
     23                             sparse_sizes=(m, n), is_sorted=False)
     24     storage = storage.coalesce(reduce=op)

~\anaconda3\lib\site-packages\torch_sparse\storage.py in __init__(self, row, rowptr, col, value, sparse_sizes, rowcount, colptr, colcount, csr2csc, csc2csr, is_sorted)
     77         if value is not None:
     78             assert value.device == col.device
---> 79             assert value.size(0) == col.size(0)
     80             value = value.contiguous()
     81 

AssertionError: 

How could I solve this?

Thank you!

[RostyslavUA]

It works, thank you!

[rusty1s]

Can you install from master again? I pushed some changes that should resolve your issues.