Datasets:

Omarrran
/

english-kashmiri-1m-translations

You need to agree to share your contact information to access this dataset

This repository is publicly accessible, but you have to accept the conditions to access its files and content.

Log in or Sign Up to review the conditions and access this dataset content.

English-Kashmiri 1M Translations

This dataset contains English → Kashmiri machine translation outputs generated in 10 splits.

Dataset Viewer

The Hugging Face Dataset Viewer is configured to read the Parquet file under:

data/train-00000-of-00001.parquet

This avoids the viewer trying to parse the XLSX artifacts directly.

Files

Requested release files:

splits/ — all 10 translated XLSX split files.
full/english_kashmiri_1m_merged.xlsx — one merged XLSX containing all split rows.
jsonl/english_kashmiri_1m.jsonl — line-delimited JSON version of the merged data.

Viewer-compatible file:

data/train-00000-of-00001.parquet — same rows in Parquet format for HF Dataset Viewer.

Additional provenance/check files:

metadata.json — row counts, schema, file hashes, and sample rows.

Row counts

Total rows: 984,764
Rows with non-empty Kashmiri translation: 984,764
Splits: 10

Schema

Fields:

id: 1-based global row id
split: source split number, 0–9
split_row_id: 1-based row number within that split
sentences: original Sentences column
source: source English text column
kashmiri: translated Kashmiri text

Notes

The source XLSX split files are preserved exactly as generated. The merged XLSX, JSONL, and Parquet were created from those split files.