Hi, this is a great little library!
I'm trying to log some reconstruction results to wandb. However, I don't see any built-in code for a testing pipeline, which means that I can't use the native wandb callback class, so if I want to manually log to wandb I have to init a new run. I want to have my reconstruction visualizations and the training stats in the same run. Is there any way to do this?
Hi, this is a great little library!
I'm trying to log some reconstruction results to wandb. However, I don't see any built-in code for a testing pipeline, which means that I can't use the native wandb callback class, so if I want to manually log to wandb I have to init a new run. I want to have my reconstruction visualizations and the training stats in the same run. Is there any way to do this?