Knodle provides an abstraction for data and models. Use it to quickly build your research benchmark.
Knodle persists several preprocessed / formatted datasets which can be used out-of-the-box.
Plug in various weak-supervision methods in combination with various prediction models, e.g. BERT.
This research is funded by the WWTF though the project “Knowledge-infused Deep Learning for Natural Language Processing” (WWTF Vienna Research Group VRG19-008).