Datasets:
Sangraha
Sangraha is the largest high-quality, cleaned Indic language pretraining data containing 251B tokens summed up over 22 languages, extracted from curated sources, existing multilingual corpora and large scale translations.
Coming Soon:
- Sangraha Synthetic - Translated and Romanised English Wikimedia data.
- Sangraha Verified - Hindi YouTube transcribed data.
More information:
- For detailed information on the curation and cleaning process of Sangraha, please checkout our paper on Arxiv;
- Check out the scraping and cleaning pipelines used to curate Sangraha on GitHub;
Getting Started
For downloading the entire Sangraha:
from datasets import load_dataset
dataset = load_dataset("ai4bharat/sangraha")
For downloading a subset (Verified/Unverified) of Sangraha:
from datasets import load_dataset
dataset = load_dataset("ai4bharat/sangraha", data_dir="<subset_name>")
# for example: dataset = load_dataset("ai4bharat/sangraha", data_dir="verified")
For downloading one language from a subset of Sangraha:
from datasets import load_dataset
dataset = load_dataset("ai4bharat/sangraha", data_dir="<subset_name>/<lang_code>")
# for example: dataset = load_dataset("ai4bharat/sangraha", data_dir="verified/asm")
Background
Sangraha contains three broad components:
- Sangraha Verified: Containing scraped data from "human-verified" Websites, OCR-extracted data from high quality Indic language PDFs, transcribed data from various Indic language videos, podcasts, movies, courses, etc.
- Sangraha Unverfied: High quality Indic language data extracted from existing multilingual corpora employing perplexity filtering using n-gram language models trained on Sangraha Verified.
- Sangraha Synthetic: WikiMedia English translated to 14 Indic languages and further "romanised" from 14 languages by transliteration to English.
Data Statistics
| Lang Code | Verified | Synthetic | Unverified | Total Tokens (in Millions) |
|---|---|---|---|---|
| asm | 292.1 | 11,696.4 | 17.5 | 12,006.0 |
| ben | 10,604.4 | 13,814.1 | 5,608.8 | 30,027.5 |
| brx | 1.5 | - | - | 1.5 |
| doi | 0.06 | - | - | 0.06 |
| eng | 12,759.9 | - | - | 12,759.9 |
| gom | 10.1 | - | - | 10.1 |
| guj | 3,647.9 | 12,934.5 | 597.0 | 17,179.4 |
| hin | 12,617.3 | 9,578.7 | 12,348.3 | 34,544.3 |
| kan | 1,778.3 | 12,087.4 | 388.8 | 14,254.5 |
| kas | 0.5 | - | - | 0.5 |
| mai | 14.6 | - | - | 14.6 |
| mal | 2,730.8 | 13,130.0 | 547.8 | 16,408.6 |
| mar | 2,827.0 | 10,816.7 | 652.1 | 14,295.8 |
| mni | 7.4 | - | - | 7.4 |
| npi | 1,822.5 | 10,588.7 | 485.5 | 12,896.7 |
| ori | 1,177.1 | 11,338.0 | 23.7 | 12,538.8 |
| pan | 1,075.3 | 9,969.6 | 136.9 | 11,181.8 |
| san | 1,329.0 | 13,553.5 | 9.8 | 14,892.3 |
| sat | 0.3 | - | - | 0.3 |
| snd | 258.2 | - | - | 258.2 |
| tam | 3,985.1 | 11,859.3 | 1,515.9 | 17,360.3 |
| urd | 3,658.1 | 9,415.8 | 1,328.2 | 14,402.1 |
| tel | 3,706.8 | 11,924.5 | 647.4 | 16,278.7 |
| Total | 64,306.1 | 162,707.9 | 24,307.7 | 251,321.0 |
To cite Sangraha, please use:
@misc{khan2024indicllmsuite,
title={IndicLLMSuite: A Blueprint for Creating Pre-training and Fine-Tuning Datasets for Indian Languages},
author={Mohammed Safi Ur Rahman Khan and Priyam Mehta and Ananth Sankar and Umashankar Kumaravelan and Sumanth Doddapaneni and Suriyaprasaad G and Varun Balan G and Sparsh Jain and Anoop Kunchukuttan and Pratyush Kumar and Raj Dabre and Mitesh M. Khapra},
year={2024},
eprint={2403.06350},
archivePrefix={arXiv},
primaryClass={cs.CL}
}