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
../
filetrain-00000-of-00001.parquet779.3 MBdownload

SPRINGLab/BPCC_cleaned · main · files are served by the source, never re-hosted here