|
|
--- |
|
|
language: en |
|
|
license: mit |
|
|
library_name: transformers |
|
|
pipeline_tag: summarization |
|
|
tags: |
|
|
- legal |
|
|
- summarization |
|
|
- nlp |
|
|
- bert |
|
|
- rag |
|
|
model_name: NayaLLM Legal Summarizer |
|
|
--- |
|
|
# NayaLLM – Legal Document Summarization Model |
|
|
|
|
|
This repository contains a Legal-BERT–based model fine-tuned for abstractive summarization of Indian legal case documents. |
|
|
|
|
|
## Model Details |
|
|
- Base Model: Legal-BERT |
|
|
- Task: Legal text summarization |
|
|
- Domain: Indian legal judgments and case documents |
|
|
|
|
|
## Pipeline Overview |
|
|
1. Legal document chunking |
|
|
2. Transformer-based embeddings |
|
|
3. Abstractive summarization |
|
|
4. Retrieval-ready outputs for RAG systems |
|
|
|
|
|
## Datasets |
|
|
1. NayaAnumana Dataset |
|
|
|
|
|
## Intended Use |
|
|
- Legal document summarization |
|
|
- Retrieval-Augmented Generation (RAG) pipelines |
|
|
- Legal research assistance |
|
|
|
|
|
## Limitations |
|
|
- Trained on domain-specific legal text |
|
|
- Not suitable for general-purpose summarization |
|
|
|
|
|
## Author |
|
|
Rakshit Gupta and Mayank Shukla |
