Affine registration and GlobalNet about deepreg HOT 4 CLOSED

Hi,

I just created a new branch with new network build. Affine transform can now be accessed while globalnet still outputs a ddf like localnet.

What do you need in terms of testing ? Should I just provide a yaml config file to use with train_and_predict_with_config just like in https://github.com/ucl-candi/DeepReg/blob/master/test/output/test_config.py

Also I noticed the labeled argument was needed for build_model() in train.py, but is missing in build_dvf_model and build_cond_model. Should I create an issue and sort it out ?

from deepreg.

It seems that the global net option has been added into the code, but tests is needed to cover it

from deepreg.

I think adding docstring to all the functions you added or related to you for one.
@NMontanaBrown can you point out where to find more information on testing to Remi?
@mathpluscode can you have a look at the potential bug mentioned?

from deepreg.

@RemiDelaunay all the data loaders only support ddf models,
i didn't have the chance to adapt dvf and conf models, but will do so this week

from deepreg.