Hi @avm21, it looks like we don’t download these on resumed runs but rather we don’t update them unless you explicitly change them on a resumed run. If you need to get group/job_type you can use the public API like this to access anything you may need:
import wandb
from wandb import Api
api = Api()
with wandb.init(entity=wandb_entity, project=wandb_project, id=run_id, resume="must") as wandb_r:
config = wandb_r.config
# A resumed run will still have the path attribute which can be used to access the run via the API
api_run = api.run(wandb_r.path)
# This will correctly print the group of the run
print(api_run.group)
Let me know if you have any questions around this.
Thank you,
Nate