Graph visualization clarification

[zoink]

I generated a graph are there a few operations that are unclear to me since I did not explicitly call them in PyTorch.

TransposeBackward - is this included with all nn.Linear operations?
AddmmBackward - is the short for add, matrix multiply for nn.Linear operations?

Below is a snippet of my graph.

[lanpa]

Yes. A linear kernel is f(x) = X * W^T + b
Here Backward is pytorch's internal representation. You can ignore it.

more information on addmm:
http://pytorch.org/docs/master/torch.html#torch.addmm

[zoink]

Thanks!