planning 2018.05.14

agenda

library name
- gradibus
- pip package and readthedocs
notebooks
- tutorial: notebook 4 (advanced concepts: caching/saving)
- tutorial: notebook 5 (Keras)
tasks
- implement make_transformer() (GŁ)
- API for ensembling (MR)
- Unintuitive adapter syntax (GŁ)
cache/save output
- Why do we have to define a caching directory for each and every transformer? Is there a better way to do it (Per project? Context handler? Something else?)
- How do we handle partitioning into training and test datasets? (just use separate data nodes?)
Input has complicated notation: nested dicts in input -> simplify interface -> DataStep should merge input_step and input_data into one API piece.

data = {'input':
          {
               'X': X_train,
               'y': y_train,
           }
        }