Software Development
I am developing a Python library called trouver which provides utilities to create and maintain mathematical Obsidian.md notes, to gather data from them to train machine learning (ML) models, and to use these models to extract and label information about such notes.
Here are ML models that I have trained; these are publicly available and open source on HuggingFace repositories:
hyunjongkimmath/information_note_type
This model classifies excerpts of mathematical text.
hyunjongkimmath/def_and_notat_token_classification_model
This model uses token classification to identify definitions and notations introduced in an excerpt of mathematical text.
hyunjongkimmath/notation_identification (deprecated)
This model identifies notations introduced in an excerpt of mathematical text. It is now deprecated in favor of the model "def_and_notat_token_classification_model".
hyunjongkimmath/notation_summarizations_model
This model summarizes what notations introduced mathematical text denote.