|
|
--- |
|
|
title: Konkani LLM |
|
|
emoji: 📊 |
|
|
colorFrom: blue |
|
|
colorTo: red |
|
|
sdk: static |
|
|
pinned: false |
|
|
license: apache-2.0 |
|
|
--- |
|
|
# Konkani LLM |
|
|
|
|
|
A suite of large language models fine-tuned to support **Konkani** in **Devanagari, Roman, and Kannada scripts**. |
|
|
These models focus on **translation, transliteration, question answering, and conversational tasks** in Konkani. |
|
|
|
|
|
## Key Features |
|
|
- Multiscript Support (Devanagari, Roman, Kannada) |
|
|
- Translation: Konkani ↔ English, Hindi, Portuguese |
|
|
- Question Answering and Dialogue |
|
|
- Script-aware Tokenization for Improved Efficiency |
|
|
|
|
|
## License |
|
|
The base model is derived from **Meta’s LLaMA / Google’s Gemma**, |
|
|
and thus this project is distributed under their respective license terms. |
|
|
Please review: |
|
|
- [LLaMA Community License](https://ai.meta.com/resources/models-and-libraries/llama-downloads/) |
|
|
- [Gemma License](https://ai.google.dev/gemma/terms) |
|
|
|
|
|
--- |
|
|
|
|
|
_Developed by Reuben for advancing AI support in low-resource Indian languages._ |
