Here, you can test the commit intent classification via the fine-tuned seBERT model.
The pre-trained version of the model can be found here.
The version that is fine tuned with the ground truth of this publication can be found here. This is also the model that classifies the text from this website.
You can either use one of our examples or provide your own commit message in the text field. The model will evaluate the commit message and return a label (other, perfective, corrective) and likelihood.