njallskarp / finetune-qa-powerset

Finetuning BERT models on a powerset of different linguistic domains

https://lvl.ru.is/

5 stars 0 forks source link

Loading model function made and refactored for more concise code #43

Closed lsig closed 1 year ago

lsig commented 1 year ago

Loading BERT models

In this PR I made a models folder with a models.py where either the user can load icebert or convbert. This code is designed to be extensible so the only change needed to add a model should be adding a model to the model dictionary.

How to test

Ruff linting with gha
No unit tests

lsig commented 1 year ago

I think the Codacy SCA does not review the code after the lint action is done, which is a problem since this should not fail since black should already have formatted the code and removed trailing whitespaces. But it seems that the lint.yml is fixed since I intentionally had a line to long to see if the lint check would fail but it did not.