Credits and licences

IndicSBERT (L3Cube)

Used for on-device matching of text across supported languages.

Model: l3cube-pune/indic-sentence-similarity-sbert, by L3Cube Pune (Deode, Gadre, Kajale, Joshi, Joshi).
Licence: CC BY 4.0 - https://creativecommons.org/licenses/by/4.0/
Source: https://huggingface.co/l3cube-pune/indic-sentence-similarity-sbert
Paper: L3Cube-IndicSBERT, arXiv:2304.11434 (https://arxiv.org/abs/2304.11434)
Changes: none to the weights. Yovumi converts them for local inference only.
Training data: trained on translated XNLI/STSb data; see licences.md.

This model runs in development evidence only and is not enabled for users.