Surpassing the much larger Gemma-2B model from Google in the 2.5B parameter category

H2O.ai, the open source leader in Generative AI and machine learning, is proud to announce that its latest open-weights (Apache v2.0) small language model, H2O-Danube2-1.8B, has secured the top position on the Hugging Face Open LLM Leaderboard for the <2B range, even surpassing the much larger Gemma-2B model from Google in the 2.5B parameter category. This achievement underscores H2O.ai's commitment to advancing AI accessibility and performance through innovative open-source solutions.

This press release features multimedia. View the full release here: https://www.businesswire.com/news/home/20240410306842/en/

Hugging Face Open LLM Leaderboard Average Scores (Graphic: Business Wire)

H2O-Danube2-1.8B is built upon the success of its predecessor, H2O-Danube 1.8B, with notable upgrades and optimizations that have propelled it to the forefront of the 2B SLM category. Leveraging a vast dataset of 2 trillion high-quality tokens, this model builds upon the Mistral architecture and optimizations, such as dropping windowing attention, to deliver unparalleled performance in natural language processing tasks.

“We love this category – a great size to fine tune or post-train on domain specific datasets for our enterprise customers, economically efficient on inference and training, and very easily embedded on edge devices like mobile phones, drones and in offline applications,” commented Sri Ambati, CEO and Founder of H2O.ai. “This model not only outperforms leading competitors like Microsoft Phi-2 and Google Gemma 2, but also provides economic efficiency and ease of deployment for enterprise and edge computing applications.

“With H2O-Danube2-1.8B, we are witnessing the democratization of large language models while ensuring sustainability within current systems,” added Sri Ambati. “The applications of this model are far reaching, from detecting and preventing PII data leakage to improving prompt generation and enhancing guardrails and the robustness of RAG systems.”

H2O.ai continues to drive innovation in AI research and development, empowering organizations to leverage cutting-edge technology without the constraints of traditional resource-intensive approaches.

Try Danube on Hugging Face.

About H2O.ai

Founded in 2012, H2O.ai is at the forefront of the AI movement to democratize Generative AI. H2O.ai’s open source Generative AI and Enterprise h2oGPTe, combined with Document AI and the award-winning autoML Driverless AI, have transformed more than 20,000 global organizations and over half of the Fortune 500 and household brands, including AT&T, Commonwealth Bank of Australia, PayPal, Chipotle, ADP, Workday, Progressive Insurance and AES. H2O.ai’s AI for Good program supports nonprofit groups, foundations and communities in their efforts to advance education, healthcare and environmental conservation, including identifying areas vulnerable to natural disasters and protecting endangered species.

H2O.ai has a vibrant community of 2 million data scientists worldwide, and aims to bring together the world’s top data scientists with customers to co-create GenAI applications that are usable and valuable by everyone. Business users can now leverage the power of LLMs to enhance productivity with enterprise applications.

View source version on businesswire.com: https://www.businesswire.com/news/home/20240410306842/en/