Documentation about what parameters can be overrides in Auto3dSeg

[KumoLiu]

Relatedly - are the full set of options for training_params documented anywhere? In my case I'm trying to prevent my data being cached during training, but more generally I'd like to know how to figure out how I can adjust other options.

CC @pwrightkcl

Originally posted by @marksgraham in #7214 (comment)

[marksgraham]

Thanks for raising this issue. We have another question - maybe we'll keep them on all this issue as the answers to all these questions should probably end up in the expanded auto3dseg documentation:

How do we force a fold to restart training from the beginning? Currently if a fold fails due to e.g. the training diverging, re-starting the runner will cause it to skip over that fold because it is internally marked as trained.

[KumoLiu]

Hi @marksgraham, AutoRunner will only consider the fold that has been trained when the metric is generated.

MONAI/monai/apps/auto3dseg/utils.py

Line 58 in 8e134b8

is_trained = best_metric is not None

You can also set train=True to always retrain all algos.

MONAI/monai/apps/auto3dseg/auto_runner.py

Line 72 in 8e134b8

train: on/off switch to run training and generate algorithm checkpoints. Defaults to None, to automatically

[Bhavay-2001]

Hi @KumoLiu, can I work on this issue if it is open for beginners?

[KumoLiu]

Hi @Bhavay-2001, thanks for your interest here. Perhaps not suitable for beginners, as you will need to be familiar with Auto3ds seg and various templates.

[Bhavay-2001]

Hi @KumoLiu, thank you for your reply. Any particular issue that you can suggest me as a beginner? Thanks

[KumoLiu]

Hi @Bhavay-2001, you can start with this ticket. #6771

[Bhavay-2001]

Hi @KumoLiu, I am reading the issue but it seems a little difficult to understand. Can you please help me understand it better.