Datasets:

agnivamaiti
/

naganlp-ner-annotated-corpus

language:
  - nag
license: mit
task_categories:
  - token-classification
tags:
  - nagamese
  - naga
  - low-resource
  - ner
  - pos
pretty_name: NagaNLP NER Annotated Corpus
size_categories:
  - n<1K
configs:
  - config_name: default
    data_files:
      - split: train
        path: train.jsonl

NagaNLP NER Annotated Corpus

This dataset contains 214 sentences annotated for Named Entity Recognition (NER) and Part-of-Speech (POS) tagging in Nagamese (Naga Pidgin).

It is part of the NagaNLP project for low-resource language documentation.

Format

The dataset was converted from CoNLL format to JSONL. Each entry contains:

tokens: List of words in the sentence.
ner_tags: List of BIO tags (e.g., B-MISC, O, B-LOC).
pos_tags: List of POS tags (e.g., NOUN, VERB).

Statistics

Sentences: 214
Tokens: ~4,800

Example

{
  "tokens": ["Etu", "retreating", "monsoon", "normally", "October"],
  "ner_tags": ["O", "O", "B-MISC", "O", "B-MISC"],
  "pos_tags": ["PRON", "VERB", "NOUN", "ADV", "PROPN"]
}