Created by: zchrissirhcz

First I like VisualDL as a visualization tool during deep learning training and really appreciate your work. As a caffe user, it helps me.

By referring different training schemes, I mean on same dataset and network but different hyperparameters. In some cases, I would like to compare between different training schemes, so hopes to choose download folder for each scheme's plotted images, thus they won't be with duplicated names, like loss.png and loss(1).png. Will this feature be supported?

My second question: for precise comparison between different training schemes, I would like to reuse the save scalar data for (my own) plotting. Is there any example for reading saved scalar data for own plot, especially for python interface?

Thanks.