IndicGLUE is a natural language understanding benchmark for Indian languages. It contains a wide
variety of tasks and covers 11 major Indian languages - as, bn, gu, hi, kn, ml, mr, or, pa, ta, te.
The Winograd Schema Challenge (Levesque et al., 2011) is a reading comprehension task
in which a system must read a sentence with a pronoun and select the referent of that pronoun from
a list of choices. The examples are manually… See the full description on the dataset page: https://huggingface.co/datasets/ai4bharat/indic_glue.
