dclm-pro-ibex-merged-260315

dclm-pro-ibex-merged-260315 Arabic translations from the IBEX parallel translation pipeline. Documents have been reconstructed from translation chunks — each row is a complete document. Dataset Details Total documents: 33,245,503 Shards: 381 (zstd compressed) Merged from: 352 chunk-level shards Created: 2026-03-17 17:26 Schema Column Type Description doc id int64 Document identifier text string Full Arabic translation (chunks merged… See the full description on the dataset page:

Source: Hugging Face Hub (SultanR/dclm-pro-ibex-merged-260315). Metadata imported from the dataset’s Hub tags.