Indian Flag
Government Of India
A-
A
A+
Home/Models/AI4Bharat - Romansetu-300M -Multilingual LLM for Indian langauges using romanization

AI4Bharat - Romansetu-300M -Multilingual LLM for Indian langauges using romanization

RomanSetu is Efficiently unlocking multilingual capabilities (Indian Languages) of Large Language Models via Romanization.

About Model

RomanSetu presents an approach that involves the continual pretraining of an English LLM, such as Llama 2, on romanized text of non-English, non-Roman script languages, followed by instruction tuning on romanized data. For continual pretraining, approximately 300 million words per language were collected and transliterated using IndicXlit with romanized datasets.

AI4Bharat - Romansetu-300M -Multilingual LLM for Indian langauges using romanization

Metadata Metadata

Llama 2 Community License Agreement

Jaavid Aktar Husain, Raj Dabre, Aswanth Kumar, Jay Gala, Thanmay Jayakumar, Ratish Puduppully, Anoop Kunchukuttan

Multilingual Model

N.A.

Open

AI4Bharat

Sector Agnostic

02/05/25 11:00:59

Nikhil Narasimhan

0

Activity Overview Activity Overview

  • Downloads0
  • Redirect 2
  • File Size 0
  • Views 246

Tags Tags

  • Multilingual
  • LLaMA2
  • Llama
  • Instruction-Tuning

License Control License Control

Llama 2 Community License Agreement

More Models from AI4Bharat More Models from AI4Bharat

AI4Bharat- 500 M - RomanSetu Multilingual Native-to-Roman Model
RomanSetu is a multilingual continual pretrained transformer model designed for transliteration across six Indic languages
Llama
Instruction-Tuning
Multilingual
LLaMA2
  • See Upvoters1
  • Downloads48
  • File Size0
  • Views836
Updated 10 month(s) ago
AI4BHARAT
View Details
share
AI4Bharat- 400 M - RomanSetu Multilingual Native-to-Roman Model
RomanSetu is a multilingual continual pretrained transformer model designed for transliteration across six Indic languages
Multilingual
LLaMA2
Llama
Instruction-Tuning
  • See Upvoters1
  • Downloads81
  • File Size0
  • Views983
Updated 10 month(s) ago
AI4BHARAT
View Details
share
AI4Bharat- Maithili - IndicConformer Automatic Speech Recognition (ASR) Model
This model takes in mono-channel audio files at a 16,000 Hz sampling rate (WAV format) and outputs the transcribed text of the speech contained in the audio.
Automatic Speech Recognition
Speech-to-Text
NLP
  • See Upvoters0
  • Downloads24
  • File Size0
  • Views643
Updated 10 month(s) ago
AI4BHARAT
View Details
share
AI4Bharat- Konkani - IndicConformer Automatic Speech Recognition (ASR) Model
Automatic Speech Recognition (ASR) model for Konkani speech recognition, processing 16,000 KHz mono WAV audio and transcribing spoken content into text
Speech-to-Text
NLP
Automatic Speech Recognition
  • See Upvoters0
  • Downloads30
  • File Size0
  • Views652
Updated 10 month(s) ago
AI4BHARAT
View Details
share
AI4Bharat- Kashmiri - IndicConformer Automatic Speech Recognition (ASR) Model
This Automatic Speech Recognition (ASR) model transcribes Kashmiri speech from 16,000 KHz mono WAV audio files into text
NLP
Speech-to-Text
Kashmiri
Automatic Speech Recognition
  • See Upvoters0
  • Downloads19
  • File Size0
  • Views680
Updated 10 month(s) ago
AI4BHARAT
View Details
share
AI4Bharat - Romansetu-200M -Multilingual LLM for Indian langauges using romanization
RomanSetu is Efficiently unlocking multilingual (Indian Languages) capabilities of Large Language Models via Romanization.
Llama
Instruction-Tuning
Multilingual
LLaMA2
  • See Upvoters0
  • Downloads6
  • File Size0
  • Views271
Updated 10 month(s) ago
AI4BHARAT
View Details
share
AI4Bharat - Romansetu-100M - Multilingual LLM for Indian langauges using romanization
RomanSetu is Efficiently unlocking multilingual (Indian Languages) capabilities of Large Language Models via Romanization.
Multilingual
LLaMA2
Llama
Instruction-Tuning
  • See Upvoters0
  • Downloads10
  • File Size0
  • Views479
Updated 10 month(s) ago
AI4BHARAT
View Details
share
AI4Bharat- Kannada - IndicConformer Automatic Speech Recognition (ASR) Model
This Kannada Automatic Speech Recognition (ASR) model transcribes 16kHz mono-channel audio into text. It utilizes a Conformer-Large architecture with 120M parameters and a hybrid CTC-RNNT decoder for high-accuracy speech recognition.
Automatic Speech Recognition
Audio Processing
NLP
  • See Upvoters0
  • Downloads22
  • File Size0
  • Views705
Updated 10 month(s) ago
AI4BHARAT
View Details
share
AI4Bharat – Romanized Path – Base to Supervised Fine-Tuning (SFT)
Romansetu model is built on base pretrained model which is supervised fine tuned on instuction-following tasks using romanized Indian languages.
Instruction-Tuning
Llama
LLaMA2
Multilingual
  • See Upvoters0
  • Downloads3
  • File Size0
  • Views198
Updated 10 month(s) ago
AI4BHARAT
View Details
share
AI4Bharat-IndicTrans2 Large-1B -English-to-Hindi (Devanagari) – : Language Translation Model
A large-scale neural machine translation (NMT) model for translating English to Hindi (Devanagari) language, leveraging 1 billion parameters for high-quality translations.
Multilingual
Transformer
cross-lingual
high-quality-translation
Large Model
low-resource-NLP
NLP
Machine Translation
  • See Upvoters0
  • Downloads33
  • File Size0
  • Views840
Updated 10 month(s) ago
AI4BHARAT
View Details
share

© 2026 - Copyright AIKosh. All rights reserved.