Create realistic use cases for the package

[sebffischer]

• [ ] think of some use-cases
  • [ ] simple use case: xgboost vs torch: no factors, no missing values
  • [ ] xgboost vs deep learning: use datasets from openml and benchmark them. with missing values and factor variables
  • [ ] proper benchmark: tuning, maybe batchmark, internal tuning
  • [ ] try to fit an image dataset:
    • [ ] using a pretrained network
    • [ ] no pretrained network.
    • [ ] with data augmentation