Closed mepster closed 6 years ago
There is a visualization example which seems to show those "orange nodes", which are pretty important, because , for example, they show activation tensor sizes which are helpful in visualizing how your network looks like.
https://github.com/szagoruyko/functional-zoo/blob/master/resnet-18-export.ipynb
However, I think that autograd changes between pytorch 0.3 and 0.4 broke that path.
Ok, there's an official answer about this question:
right, not possible without internal changes to pytorch, closing.
In dot.py there is a case in which a node will be plotted with fillcolor='orange'. However, none of the example networks in examples.ipynb show orange boxes - only blue and gray. Under what conditions are you supposed to get orange boxes? Can you tweak one of the example networks to illustrate this?
I've been playing with the code in dot.py, but I haven't been able to define a network that produces orange boxes... I don't think "torch.is_tensor(var)" ever evaluates to true... at least in no network I could come up with. (However, "torch.is_tensor(var.variable)" does sometimes evaluate true.)
Thanks! (Very useful tool!)