language:
- pi
- my
license: mit
task_categories:
- translation
- text-generation
- question-answering
tags:
- pali
- myanmar
- dictionary
- buddhist
- nlp
- linguistics
- instruction-finetuning
dataset_info:
features:
- name: word
dtype: string
- name: part_of_speech
dtype: string
- name: etymology
dtype: string
- name: definition
dtype: string
- name: instruction
dtype: string
- name: response
dtype: string
splits:
- name: train
num_examples: 306063
Pali-Myanmar Dictionary Corpus (Instruction-Ready)
Dataset Summary
The Pali-Myanmar Dictionary Corpus is an extensive, highly structured linguistic resource containing 306,063 entries. It serves as a comprehensive bridge between the ancient Pali language and Modern Myanmar (Burmese). This dataset is specifically designed for Natural Language Processing (NLP), Machine Translation, and Large Language Model (LLM) instruction tuning.
Each record is parsed from original sources into a structured format that includes grammatical categories, etymological roots, and clean definitions.
Project Credits & Governance
- Organization: DatarrX
- Curated by: Khant Sint Heinn (Kalix Louis)
- Source Data: Derived from pndaza/tipitaka-pali
- License: MIT License
Understanding the Data (For Non-Religious Researchers)
Pali is a classical Indo-Aryan language, famous as the language of the Buddhist scriptures (Tipitaka). However, this dataset is more than a religious text; it is a Classical-to-Modern Language Mapping Corpus.
For AI researchers, this dataset is valuable for:
- Morphological Analysis: Pali has a complex inflectional system similar to Latin or Sanskrit.
- Low-Resource NLP: Providing high-quality supervised data for the Myanmar script.
- Semantic Mapping: Understanding how ancient philosophical concepts translate into modern legal and ethical Burmese terminology.
Dataset Statistics
| Metric | Value |
|---|---|
| Total Entries | 306,063 |
| Average Word Length | 9.24 characters |
| Entries with Part of Speech (POS) | 232,186 (75.8%) |
| Entries with Etymology | 178,316 (58.2%) |
Top 10 Part of Speech (POS) Distribution
| POS Tag | Count | Percentage | Description |
|---|---|---|---|
| တိ | 64,733 | 21.15% | Adjectives / Pronouns |
| န | 41,840 | 13.67% | Neuter Nouns |
| ပု | 37,432 | 12.23% | Masculine Nouns |
| ထီ | 15,715 | 5.13% | Feminine Nouns |
| ပ | 12,520 | 4.09% | Prefixes / Primary Roots |
| ကြိ | 10,574 | 3.45% | Verbs |
Dataset Structure
The dataset follows an Instruction-Response format, making it compatible with modern LLM training (Supervised Fine-Tuning).
Data Fields
word: The target Pali word.part_of_speech: Grammatical classification (Gender, Tense, etc.).etymology: The root origin of the word.definition: The modern Myanmar translation/definition.instruction: A natural language prompt for the AI.response: A human-like generated response combining grammar and definition.
Example Entry
{
"word": "စက္ကဒဟ",
"part_of_speech": "ပု",
"etymology": "စက္က+ဒဟ",
"definition": "စကြာရတနာဖြစ်ပေါ်ရာအရပ်။",
"instruction": "Provide the definition and grammar for the Pali word: စက္ကဒဟ",
"response": "စက္ကဒဟ (ပု) is defined as: စကြာရတနာဖြစ်ပေါ်ရာအရပ်။"
}
Potential Applications
- Instruction Tuning for LLMs: Fine-tuning models like Llama or Gemma to understand and explain Pali terms.
- Machine Translation (MT): Creating a baseline vocabulary for Pali-to-Burmese translation systems.
- Automatic POS Tagging: Using the 232k+ POS labels to train grammatical classifiers for the Myanmar script.
- Digital Humanities: Researching the linguistic evolution and word-formation (etymology) of Pali words in the Myanmar context.
Citation & Acknowledgments
If you use this corpus, please acknowledge the source and the developers:
Source Data: Pndaza (2024). Tipitaka Pali SQLite Database. Available at: https://github.com/pndaza/tipitaka-pali Curation: Khant Sint Heinn (Kalix Louis), DatarrX Organization (2026). Pali-Myanmar Dictionary Corpus.
Contact
For inquiries regarding DatarrX's Myanmar NLP initiatives, please visit Hugging Face or our GitHub.