lanpa/tensorboardX

Multiple scalars in the same plot, without separate runs

zimonitrome opened this issue · 1 comments

Is this possible in TensorboardX? I have only used Pytorch's SummaryWriter.

Their solution is to use add_scalars but this creates multiple "runs" with different colors. Ideally these combined scalars would belong to a single run while still be viewed in the same plot/graph perhaps with the same color but different line thickness/dashing.

I really like using Tensorboard but if someone knows a better package for my use-case, please let me know.

Same problem here