CoolFace
Datasetpublic

SPRINGLab/BPCC_cleaned

A curated subset of Bharat Parallel Corpus Collection (BPCC) for 8 Indian languages. Translation pairs are filtered with LABSE score(>0.9) and further preprocessed. Useful for training high-quality translation models.

sourceHugging Faceupdated 2y agoView on Hugging Face
1likes502downloads
Dataset Card

A curated subset of Bharat Parallel Corpus Collection (BPCC) for 8 Indian languages. Translation pairs are filtered with LABSE score(>0.9) and further preprocessed. Useful for training high-quality translation models.