- GenAI Foundry
GenAI Foundry
The enterprise GenAI platform for full control over your model, data, and intelligence — tailored for regulated industries.
Build, Fine-Tune & Deploy Private GenAI Models Securely
- InsurancGPT
AI Purpose-Built for the Insurance Industry
InsurancGPT
Custom-tuned suite of LLMs trained on deep insurance domain data including P&C, Auto, Health, and Life
The Core Intelligence Engine for Insurance AI
- NammaKannadaGPT
NammaKannadaGPT
Foundational Large Language models for native languages
- ROI Calculator
ROI Calculator
Transforming Business Efficiency with the Enkefalos ROI Calculator
- Guardian
Enkefalos Guardian
Your Control Center for Responsible AI in Insurance
NammaKannadaGPT
Foundational Large Language models for native languages
ನಮ್ಮ ಕನ್ನಡ ಜಿಪಿಟಿ
ಎನ್ಕೆಫಲೋಸ್ ಟೆಕ್ನಾಲಜೀಸ್ನಲ್ಲಿ ನಾವು ನಮ್ಮ ಕನ್ನಡ ಜಿಪಿಟಿಯನ್ನು ಅಭಿವೃದ್ಧಿಪಡಿಸಿದ್ದೇವೆ, ಇದು ಕನ್ನಡಕ್ಕಾಗಿ ಮೊದಲಿನಿಂದ ನಿರ್ಮಿಸಲಾದ ಫೌಂಡೇಶನ್ ಲಾರ್ಜ್ ಲ್ಯಾಂಗ್ವೇಜ್ ಮಾಡೆಲ್ (ಎಲ್ಎಲ್ಎಂ) ಆಗಿದೆ. ಅಸ್ತಿತ್ವದಲ್ಲಿರುವ ವಾಸ್ತುಶಿಲ್ಪಗಳನ್ನು ಆಧರಿಸಿದ ಫೈನ್-ಟ್ಯೂನ್ ಮಾಡಲಾದ ಮಾದರಿಗಳಿಗಿಂತ ಭಿನ್ನವಾಗಿ, ನಮ್ಮ ಕನ್ನಡ ಜಿಪಿಟಿಯನ್ನು ಕನ್ನಡ ಪಠ್ಯದ ಮೇಲೆ ಪ್ರತ್ಯೇಕವಾಗಿ ತರಬೇತಿ ನೀಡಲಾಗಿದೆ.
At Enkefalos Technologies, we have developed our KannadaGPT, a foundational Large Language Model (LLM) built from scratch for Kannada. Unlike fine-tuned models based on existing architectures, our KannadaGPT is trained exclusively on Kannada text.
This model marks a significant step towards enabling AI that understands and communicates in Kannada. While not yet highly accurate, our Kannada GPT lays the foundation for future advances in Kannada AI.
As part of our commitment to Kannada AI and the open-source community, we are open-sourcing this model to foster collaboration and continuous improvement. The model is available in Hugging Face, as well as an interactive ChatGPT-like interface for users and developers. We invite AI researchers, organizations, and community contributors to explore, refine, and help us improve the model to ensure Kannada AI continues to evolve and thrive
NammaKannadaGPT Technical Information
Our Kannada GPT is a GPT transformer with 162.42 million trainable parameters. It was trained on 330.5 million tokens using a Kannada tokenizer with a context length of 256. The training process consisted of 290,508 steps in 76 hours, using high-performance hardware (24-core GPU, 220 GB RAM, 64 GB disk).
Areas of use of our Kannada GPT
See how our Kannada GPT can help you boost your performance.
Research and content creation
- Creation of articles, essays and blog posts
- Summarizing research papers and documents
- Creation of creative Kannada poetry and stories
Education and learning
- Kannada grammar and writing help
- Kannada translation of textbooks and study materials
- Kannada language learning and vocabulary help
Business and productivity
- Professional email and document creation
- Preparation of reports, business plans and proposals
- Preparing Kannada answers for customer support
Media and Entertainment
- Kannada script writing for films, plays and dramas
- Subtitles and audio recording script preparation
- Social media content creation in Kannada
AI chatbots and virtual assistants
- Implementation of Kannada chatbots for businesses
- Improving communication with customers in Kannada
- Providing FAQ and automated answers in Kannada
Legal and government services
- Preliminary Drafting of Kannada Legal Documents
- Kannada translation of policies and regulations and management of government publications
- Legal advice and fulfillment of applications in Kannada
NammaKannadaGPT
Foundational Large Language models for native languages