This dataset contains 108,463 human-labeled and 656k noisily labeled pairs that feature the importance of modeling structure, context, and word order information for the problem of paraphrase identification.
Primary LanguagePythonOtherNOASSERTION
No one’s watching this repository yet.