README.md

metadata

license: mit
task_categories:
  - text-classification
language:
  - asm
  - ben
  - brx
  - doi
  - gom
  - guj
  - hin
  - kan
  - kas
  - mai
  - mal
  - mar
  - mni
  - npi
  - ory
  - pan
  - san
  - sat
  - snd
  - tam
  - tel
  - urd
  - eng
tags:
  - language-identification
  - indian-languages
pretty_name: ILID
size_categories:
  - 100K<n<1M

ILID: Native Script Language Identification for Indian Languages

Paper | Code | Project Page

🗣 ILID: Indian Language Identification Dataset (23 Languages)
Authors: Yash Ingle, Dr. Pruthwik Mishra
Institute: Sardar Vallabhbhai National Institute of Technology (SVNIT), Surat, India

---

📄 Dataset Description

The ILID (Indian Language Identification Dataset) benchmark contains 250,000 sentences from English and 22 official Indian languages, designed for training and evaluating language identification models. The dataset supports the task of distinguishing between Indian languages, many of which share scripts, vocabulary, and structure.

---

📊 Dataset Statistics

Language	Code	Train	Dev	Test	Total
Assamese	asm	8000	1000	1000	10000
Bengali	ben	8000	1000	1000	10000
Bodo	brx	8000	1000	1000	10000
Dogri	doi	8000	1000	1000	10000
Konkani	gom	8000	1000	1000	10000
Gujarati	guj	8000	1000	1000	10000
Hindi	hin	8000	1000	1000	10000
Kannada	kan	8000	1000	1000	10000
Kashmiri	kas	8000	1000	1000	10000
Maithili	mai	8000	1000	1000	10000
Malayalam	mal	8000	1000	1000	10000
Marathi	mar	8000	1000	1000	10000
Manipuri (Bengali)	mni_Beng	8000	1000	1000	10000
Manipuri (Meitei)	mni_Mtei	8000	1000	1000	10000
Nepali	npi	8000	1000	1000	10000
Odia	ory	8000	1000	1000	10000
Punjabi	pan	8000	1000	1000	10000
Sanskrit	san	8000	1000	1000	10000
Santali	sat	8000	1000	1000	10000
Sindhi (Arabic)	snd_Arab	8000	1000	1000	10000
Sindhi (Devanagari)	snd_Deva	8000	1000	1000	10000
Tamil	tam	8000	1000	1000	10000
Telugu	tel	8000	1000	1000	10000
Urdu	urd	8000	1000	1000	10000
English	eng	8000	1000	1000	10000
Total	—	200000	25000	25000	250000

---

📁 Files Provided

• shuffled_train_sentences: Training sentences (80% split – 200,000 samples)
• shuffled_train_labels: Corresponding labels for training sentences
• shuffled_dev_sentences: Validation (dev) sentences (10% split – 25,000 samples)
• shuffled_dev_labels: Corresponding labels for dev sentences
• shuffled_test_sentences: Test sentences (10% split – 25,000 samples)
• shuffled_test_labels: Corresponding labels for test sentences

📌 Tasks

• Language Identification (LID)
• Multilingual Text Classification
• Benchmarking ML & DL Models on Indian Languages

---

🧹 Data Collection & Cleaning

• 13 languages collected using web scraping from Wikipedia, news portals, and blogs.
• 10 languages sampled from large monolingual corpora (Bhashaverse).
• Each sentence underwent cleaning, normalization, and language filtering via FastText.

---

🧠 Models & Results

Baseline models include:

• TF-IDF + Machine Learning: SVM, Logistic Regression, Random Forest, etc.
• FastText Classifier
• Fine-tuned MuRIL (BERT for Indian languages)

Best ensemble models achieve F1-scores of up to 0.99 on test/dev sets.

---

📚 Citation

If you use this dataset, please cite:

@misc{ingle2025ilidnativescriptlanguage,
      title={ILID: Native Script Language Identification for Indian Languages}, 
      author={Yash Ingle and Pruthwik Mishra},
      year={2025},
      eprint={2507.11832},
      archivePrefix={arXiv},
      primaryClass={cs.CL},
      url={https://arxiv.org/abs/2507.11832}, 
}