Daisy-chaining graphs of multiple checkpointed runs

[patocal]

Hi, I’m super impressed with wandb’s ability to interact with our local cluster.
I’m in the process of writing up a project and I just need to tidy up multiple runs that were restarted from checkpoints. The image below should give you an idea of my issue. Pointers to the docs would be helpful. I appreciate I can use pandas to do this using the underlying data, but was wondering if there is a direct way.

In a related vein, when I copy a directory and start a fresh run there (so as to preserve the original), I find that the old ones don’t appear (as a new user I can’t post a photo).

[artsiom]

Hi @patocal!

Thank you for writing in!
To confirm are you interested in having all of those runs show up at the single color of a single run?

Warmly,
Artsiom

[artsiom]

Hi Patrick,

We wanted to follow up with you regarding your support request as we have not heard back from you. Please let us know if we can be of further assistance or if your issue has been resolved.

Best,
Weights & Biases

[patocal]

Thanks Artsiom: I got by without doing that. I have a more important problem that I’ll request help for in a new post.

[artsiom]

Sounds great! Thank you for your follow up and looking forward to hearing from you!