Discover Sherlock, Poland’s first generative AI platform by CloudFerro. Featuring Bielik 11B v2.3 and Llama 3.1 70B, it offers advanced, localized AI solutions while ensuring data security and compliance.
CloudFerro Introduces Sherlock: A Cutting-Edge Generative AI Platform Built for Local Needs
CloudFerro, the leading provider of Poland’s largest public cloud, has unveiled Sherlock, a groundbreaking platform offering access to advanced generative AI models. Designed to meet local needs, Sherlock empowers organizations to leverage artificial intelligence without the complexity of managing infrastructure, ensuring full control over data and compliance with local regulations.
At the heart of the platform lies Bielik 11B v2.3 Instruct, Poland’s first large-scale language model trained on Polish datasets, created in collaboration with SpeakLeash and Cyfronet AGH. This model delivers unparalleled proficiency in handling Polish-language documents, understanding cultural nuances, and providing context-aware insights. Complementing it is Llama 3.1 70B, an advanced multilingual model tailored for complex global tasks. Both models are accessible via a standard OpenAI-compatible API, enabling seamless integration into various applications.
Advanced Text Embeddings and Semantic Analysis
Sherlock also provides access to powerful text embedding models, including BGE-Multilingual-Gemma2 and e5-mistral-7b-instruct. These tools facilitate semantic search and large-scale data analysis, supporting industries such as law, medicine, and science. They enable the creation of sophisticated chatbots, recommendation systems, and document processing solutions tailored to specific needs.
Revolutionizing Earth Observation with AI
In partnership with ESA’s Φ-lab, CloudFerro has developed global AI embeddings for Earth observation, transforming satellite data into vector-based numerical representations. Leveraging over 62 TB of Copernicus data and models like DINOv2, the initiative produced 200 million embeddings, streamlining workflows for remote sensing and geospatial analytics.
Empowering Local AI Innovation
Sherlock not only integrates global models but also drives the development of Polish AI solutions. With a cloud architecture powered by high-performance GPUs, the platform lowers the barriers to AI adoption, making advanced technology accessible even to smaller organizations.
CloudFerro emphasizes the importance of localized AI solutions to address language and cultural nuances, as well as stringent data security requirements. “Our models operate in a closed environment, ensuring user data is neither stored nor used for training. This is crucial for regulated sectors like healthcare, public administration, and finance,” said Jan Szypulski, Product Manager at CloudFerro.
Future Developments for Sherlock
CloudFerro plans to enhance Sherlock with additional language models, image generation capabilities, and expanded Earth observation tools, addressing the evolving needs of AI users while maintaining a user-friendly experience.
Sherlock exemplifies the potential for decentralized, locally driven AI innovation, offering unique value to Polish organizations while competing with global players