Team Oasis colab notebooks and stuff for National Tech Olympycs - 2021
Typical text classification
-
models used:
-
Dataset
There are a lot of ways to handle this dataset. We tried (partially succesfull): QA, Classification, Span generation with and withous NER
-
Dataset
-
models (transforsmers only, obviously) used:
- For text classification: https://huggingface.co/DeepPavlov/rubert-base-cased ,via slavic langs vocab provided by from DeepPavlov
- For correct span generation: https://huggingface.co/google/mt5-large
- For QA: https://huggingface.co/bert-base-multilingual-cased