---
license: mit
---

**CTMatch Dataset**

This is a set of 2 labelled datasets of topic (patient descriptions), doc (clinical trials documents - mostly just extracted eligbility criteria), and label ({0, 1, 2}) triples, in json format. 

These have been processed using ctproc, and in this state can be used by various tokenizers for fine-tuning (see ctmatch for examples).