Need to refactor core methods, written in a functional fashion, into a separate class

domenicodigangi / DynWGraphsPaper

Code for analysis of paper about dynmical sparse weighted graphs

MIT License

0 stars 1 forks source link

Need to refactor core methods, written in a functional fashion, into a separate class #5

Closed domenicodigangi closed 3 years ago

domenicodigangi commented 3 years ago

Then models should have this class as superclass or attribute (probably attribute since they should inherit from torch.nn.module). And they should have parameters as attributes. Methods to be refactored:

[x] regr_product
[x] cond_exp
[x] identify
[x] start_phi_from_obs
[x] dist_from_params
[x] score ?
[x] link_dist_par
[x] dist_par_un_start_val
[ ]

domenicodigangi commented 3 years ago

done using inheritance, grouped methods in dirSpW1_funs. Sequence models inherit from dirSpW1_funs that in turn inherits from torch.nn.Module

domenicodigangi commented 3 years ago

Done for weighted fitness model