pos_weight should be a Tensor?

[shionhonda]

When I train with Cora dataset, I get the following error in binary_cross_entropy_with_logits. Shouldn't pos_weight be a Tensor? Thanks!

Traceback (most recent call last):
  File "train.py", line 83, in <module>
    gae_for(args)
  File "train.py", line 62, in gae_for
    norm=norm, pos_weight=pos_weight)
  File "/gae-pytorch/gae/optimizer.py", line 7, in loss_function
    cost = norm * F.binary_cross_entropy_with_logits(preds, labels, pos_weight=pos_weight)
  File "/anaconda3/lib/python3.6/site-packages/torch/nn/functional.py", line 2077, in binary_cross_entropy_with_logits
    return torch.binary_cross_entropy_with_logits(input, target, weight, pos_weight, reduction_enum)
TypeError: binary_cross_entropy_with_logits(): argument 'pos_weight' (position 4) must be Tensor, not numpy.float64

[shionhonda]

I fix it by replacing this:

https://github.com/zfjsail/gae-pytorch/blob/22c9edfd37d7679ebff32e4ad92eb89cd96837da/gae/train.py#L48

to this:

pos_weight = torch.Tensor([float(adj.shape[0] * adj.shape[0] - adj.sum()) / adj.sum()])

[iamadog3333]

It works.

but I found that length of pos_weight is 1.
According documents: pos_weight (Tensor, optional) – a weight of positive examples. Must be a vector with length equal to the number of classes.
pos_weight 's length should be 2.

[iamadog3333]

Doc says: where c is the class number (c>1 for multi-label binary classification, c=1 for single-label binary classification) That is right.

[Dzhilin]

Doc says: where c is the class number (c>1 for multi-label binary classification, c=1 for single-label binary classification) Hi, I have the same problem. Has your problem been solved？