nyanp
This PR resolves #3556 by implementing the following method in `CVBooster`: - save_model - model_from_string - model_to_string
Unlike the current implementation of cv in nyaggle, The models trained in lgb.cv and xgb.cv have an equal number of trees in all folds. Since these “balanced” models may work...
It would be nice if run_experiment accepts dictionary to customize layout of logging artifacts
- Kolmogorov-Smirnov test - Adversarial validation - Percentage of nulls in train/test - Unseen category in test
(update on 8/30) ## v1.0.0alpha on 12 September, 2016 - [ ] Add model serialization API @nyanp - [x] Add protobuf parser for tensorflow @Wangyida - [x] Fix compiler warnings...