promach
(buttercutter)
January 19, 2022, 3:54pm
1
I am trying to trace the tensorboard graph for GitHub - buttercutter/gdas: A simple implementation for the GDAS paper : Searching for A Robust Neural Architecture in Four GPU Hours
However, from what I can observe so far, the tensorboard graph generated from torch.utils.tensorboard does not really indicate user-understandable node names and cell names which makes it so difficult for tracking down the connections within the graph.
Any suggestions ?
1 Like
Advitya17
(Advitya Gemawat)
February 16, 2022, 2:59pm
2
promach:
However, from what I can observe so far, the tensorboard graph generated from torch.utils.tensorboard does not really indicate user-understandable node names and cell names which makes it so difficult for tracking down the connections within the graph.
I have the same question.
promach
(buttercutter)
February 19, 2022, 4:47pm
3
promach
(buttercutter)
February 20, 2022, 1:58am
4
I tried from torch.utils.tensorboard import SummaryWriter and from tensorboardX import SummaryWriter , but both are still giving some non-user-readable nodes and cells names for the output graph.
I suspect the issue might be due to the multiple nested for
loops inside the forward function .
Note: I have removed for epoch in range(NUM_EPOCHS):
during testing and debugging this issue.
Any update on this? Iām having the same issue using the EmbeddingBag
layer.