skip ml model

[gitlzzz]

What if a training was failed and skipped, it will be retraining?

https://github.com/pol-sb/MatDBForge/blob/5dcc5be7c237ae4d1e4c5dfa6ec1c66e3b68924d/src/MatDBForge/workflows/simple_active_learning.py#L350

[pol-sb]

Hey @gitlzzz,

No, there won't be any retraining. If the TrainMACEModelCalculation CalcJob finishes incorrectly, it will be skipped for the current iteration. The rest will be used.

What happens when everything fails? You get an error and the loop stops. I think that makes sense, because if the trainings cannot finish then it means there's something terribly wrong with the current dataset.

What do you think?