ZamAI Pashto Multimodal Datasets
Collection
Pashto text, voice, vision, and OCR/document datasets for open AI research. • 7 items • Updated • 1
document_id stringclasses 3
values | doc_type stringclasses 3
values | domain stringclasses 3
values | source_language stringclasses 2
values | target_language stringclasses 3
values | ocr_text stringclasses 3
values | translation stringclasses 3
values | summary stringclasses 3
values | quality_score float64 0.88 0.93 | scan_path stringclasses 3
values |
|---|---|---|---|---|---|---|---|---|---|
doc-001 | letter | government | ps | en | د کابل ښاروالۍ خبرتیا | Kabul municipality notice | Public notice about office hours. | 0.93 | data/scanned_docs/doc-001.png |
doc-002 | form | education | ps | fa | د زده کوونکي نوم | نام دانش آموز | Student registration form sample. | 0.88 | data/scanned_docs/doc-002.png |
doc-003 | receipt | commerce | fa | ps | رسید پرداخت | د تادیې رسید | Payment receipt example for OCR cleanup. | 0.9 | data/scanned_docs/doc-003.png |
Pashto
Languages: ps
License: cc-by-4.0
Task categories: visual-document-retrieval
Size categories: n<1K
Summary
This dataset is part of the ZamAI Pashto data collection. It is intended for visual-document-retrieval tasks in Pashto.
How to use
from datasets import load_dataset
dataset = load_dataset("tasal9/zamai-pashto-documents")
print(dataset)
Citation
@misc{zamai_pashto_data,
title = {{Pashto}},
author = {ZamAI / Yaqoob Tasal},
year = {2024},
howpublished = {\url{https://huggingface.co/datasets/tasal9/zamai-pashto-documents}}
}
License
This dataset is released under the cc-by-4.0 license.
- Downloads last month
- 33