A tri-lingual (English en, Odia or, Sanskrit sa) question–answer
corpus derived from the ODEN initiative.
Total rows: 717
Splits: train / validation / test = 645 / 35 / 37
License: cc-by-4.0
{
"id": "uuid",
"lang": "en|od|sa",
"question": "…",
"answer": "…"
}
from datasets import load_dataset
ds = load_dataset("BBSRguy/ODEN-Indictext", split="train")
print(ds[0])
@misc{oden-indictext-2025… See the full description on the dataset page:
https://huggingface.co/datasets/BBSRguy/ODEN-Indictext.
