HypEval is a small package to evaluate the quality of a model on the hypernymy detection task. It provides a simple and efficient interface for the evaluation on some common hypernymy datasets.
As of now, the list of datasets include:
- Baroni2012 (Entailment above the word level in distributional semantics, 2012)
- BLESS (How we BLESSed distributional semantic evaluation, 2011)
- EVALution (Evalution 1.0: an evolving semantic dataset for training and evaluation of distributional semantic models, 2015)
- Kotlerman2010 (Directional distributional similarity for lexical inference, 2010)
- LenciBenotto (Distributional Models for Semantic Relations: A Study on Hyponymy and Antonymy, 2015)
- Levy2014 (Focused entailment graphs for Open IE propositions, 2014)
- Turney2014 ( Experiments with Three Approaches to Recognizing Lexical Entailment, 2014)
- Weeds (Learning to Distinguish Hypernyms and Co-Hyponyms, 2014)
- BIBLESS (Exploiting image generality for lexical entailment detection, 2015)
- HypeNet (Improving Hypernymy Detection with an Integrated Path-based and Distributional Method, 2016)