Training model outside of EventStorage

[flgau]

I am wondering if their is a way to train a detectron2 model outside of EventStorage context
For exemple, I have already my own way of storing training information of a pytorch. I would like to be able to simply train like this :
https://pytorch.org/tutorials/beginner/blitz/cifar10_tutorial.html#train-the-network, instead of having to train in EventStorage. Is it possible? If so how can I do it? I am building my model with build_model(cfg) and same thing for optimizer :build_optimizer(model, cfg)