UserWarning: Using a target size

[anthonysirico]

Should I be concerned about the following warning?

C:\Users\Anthony Sirico\AppData\Local\Temp\ipykernel_121128\2205290571.py:10: UserWarning: Using a target size (torch.Size([1])) that is different to the input size (torch.Size([1, 1])). This will likely lead to incorrect results due to broadcasting. Please ensure they have the same size.
  loss = criterion(out, data.y)  # Compute the loss.

Because my results are coming out like this:

Epoch: 001, Train Acc: 0.0000, Test Acc: 0.0000
Epoch: 002, Train Acc: 0.0000, Test Acc: 0.0000
Epoch: 003, Train Acc: 0.0000, Test Acc: 0.0000

[rusty1s]

I think it is good practice to ensure that shapes match and that your training does not yield any warnings. So out = out.view(-1) should fix that.

Nonetheless, your low performance seems to be induced by something different.

[anthonysirico]

Do you have any examples doing any form of regression whether linear of logistic on graphs?

[rusty1s]

Here is an example of a graph-level regression task. In general, you just need to swap out your loss function and train via, e.g., torch.nn.functional.mse_loss.