-
Adding possibility to evaluate classes the freshly trained tagger is not trained on 4 of 4 checklist items completed
-
Add script that allows to add model predictions to h5 files (i.e. to training/validation/test files) 4 of 4 checklist items completed!637
-
Improve support for dropout in DIPS 4 of 4 checklist items completed
-
removing `#!/usr/bin/env python` from scripts 3 of 3 checklist items completed
-
Fix dropout in DL1* models 5 of 5 checklist items completed!633
-
Remove preprocessing config from loading functions 4 of 4 checklist items completed
-
Apply naming scheme for WP and nEpochs 4 of 4 checklist items completed
-
Adding correct naming scheme for train config sections 4 of 4 checklist items completed
-
More robust preprocessing configuration 4 of 4 checklist items completed
-
Merging load validation data functions 4 of 4 checklist items completed
-
Preprocess configs cleanup 4 of 4 checklist items completed
-
Removing var_dict from train config 3 of 3 checklist items completed
-
Splitting configuration into base class and some preprocessing adaptions 4 of 4 checklist items completed
-
Unify scaling/shifting application for preprocessing/validation 3 of 3 checklist items completed
-
Resolve "high level function to load scaled tracks and jets from h5 file" 4 of 4 checklist items completed
-
Adding metadata information to training file 4 of 4 checklist items completed
-
Compare class sels 4 of 4 checklist items completed
-
Fixing unit tests pylint 4 of 4 checklist items completed
-
Fixing lazy formatting 3 of 3 checklist items completed
-
Adding consistent n_jets naming 4 of 4 checklist items completed