Supercomputing environments are great for generating large amounts of data. However, making sense of that data can often be a mix of science and art. I want to start a thread for tools for high-dimensional data visualization!
I’m a big fan of parallel coordinate plots. For publishable-quality static versions, I like to use Paxplot. For interactive versions, I like HiPlot.
A note of disclosure, I am the creator of Paxplot and am working on a contribution to HiPlot.
Hi Jacob,
I saw on your example plot epoch name mentioned i try to compile and run. Are you running this code on Summit? Can you post your sbatch run script ? That would save me time from experimenting. Is there any small scale cluster with just couple nodes where we can use our codes for debugging and running from terminals without any sbatching?