GenVR Research unveils multi-lingual text-to-image AI model with 15 Indian languages

Image for representational purposes only. Photo Courtesy: iStock

In an ambitious leap forward for AI and cultural representation, GenVR Research has announced the launch of Rupayan, a pioneering text-to-image model tailored specifically for the Indian context. This innovative platform is designed to generate images that reflect the rich diversity of India's cultural and linguistic heritage, supporting an impressive array of 15 Indian languages, including the classical language of Sanskrit.

Rupayan stands out as a beacon of inclusivity in the technology world, aimed at bridging the gap between AI advancements and regional diversity. This model is not just a technical achievement; it's a cultural milestone that promises to make AI more accessible and relevant to millions of users across India.

Released on 24th January 2024, the company is among the first few in India to launch an indigenous text-to-image model. The platform, Matisse, where the model is hosted has now seen over 15,000 image generations.

One of the most notable features of Rupayan is its multilingual capabilities. By offering support for a wide range of Indian languages, from Hindi, Bengali, and Tamil to lesser-served languages like Sanskrit, Rupayan is set to revolutionize how content is created and consumed in the region. This inclusivity ensures that users can interact with the model in their native language, making AI a tool for all, not just the English-speaking elite.

Rupayan is specifically designed to understand and generate images that resonate with Indian aesthetics, landmarks, cultural symbols, and everyday scenes. This focus on local relevance is a significant departure from the one-size-fits-all approach seen in most global AI models, which often lack the nuance and context needed to accurately represent non-Western cultures.

Fine-tuning open source diffusion transformers based models (the key tech behind OpenAI's SORA) on highly aesthetic images, the company has solved text generation and aesthetic abilities in one go and gives competition to players like Midjourney on image quality.

The development of Rupayan underscores GenVR Research's commitment to fostering innovation that is inclusive and reflective of the world's diversity.

GenVR Research has open-sourced over 350 Indic image LORA (Low Rank Adaptation) models, 65 per cent of all open-sourced text-to-image LORA models coming out of India. This decision to share their advanced models with the public is a game-changer, encouraging further research, development, and innovation in the field of AI across the globe. By making these resources available, GenVR ensures that the benefits of such advancements are widely accessible, fostering a community of collaboration and shared progress.

This story has been sourced from a third party syndicated feed, agencies. Mid-day accepts no responsibility or liability for its dependability, trustworthiness, reliability and data of the text. Mid-day management/mid-day.com reserves the sole right to alter, delete or remove (without notice) the content in its absolute discretion for any reason whatsoever.