Post process + classification bug (ppdebreuck#53)

* initial postprocess implementation

* classif bug fix

* refit update

* pre-commit run

* v0.1.10 release

* using model.predict instead of tensor

version update

Merge pull request ppdebreuck#22 from ppdebreuck/ml-evs/readd_cross_nmi

Add ability to load precomputed cross NMI from figshare

Version update 0.1.7

refactor changes

refactor changes

Fixed various bugs,

- Make target names and material_ids more flexible by including any iterable
- Fixed target construction issue in MODNet, now expects a list or (n_materials,n_target) nd array
- Fixed evaluation of dataframes in if constructions
- Removed mpids attribute
- Fixed feature_selection .loc error
It is now possible to pass any Iterable for target names and material ids.

zero-range-features bug fix

Added large files (db and moddatas)

update readme